A retrospective evaluation of second-line chemotherapy response in hormone-refractory prostate carcinoma

Second-line taxane-based therapy after first-line epothilone-B analog ixabepilone (BMS-247550) therapy




Epothilones and taxanes interfere with microtubule function. Ixabepilone, which is an epothilone-B analog, has activity against taxane-resistant cell lines and as first-line therapy for men with hormone-refractory prostate carcinoma (HRPC). Clinical cross-resistance of ixabepilone and taxanes in HRPC is unknown.


Records were evaluated retrospectively from patients with HRPC who were treated on a randomized Phase II trial of ixabepilone with or without estramustine and who subsequently received taxane chemotherapy. Posttherapy declines in prostate-specific antigen (PSA) levels and time to PSA progression were defined by consensus criteria. The median survival was evaluated by using the Kaplan–Meier method.


Forty-nine patients who received ixabepilone with estramustine (28 patients) or without estramustine (21 patients) subsequently received second-line taxane therapy. Second-line PSA declines ≥ 50% were achieved by 51% of patients (95% confidence interval [95% CI], 33–66%). Second-line PSA declines ≥ 50% were achieved by 61% of patients (95% CI, 42–78%) who achieved a first-line PSA decline ≥ 50% with ixabepilone, compared with 33% of patients (95% CI, 13–59%) who did not (P = 0.08). Patients who discontinued first-line ixabepilone treatment for disease progression were less likely to achieve a PSA decline ≥ 50% in response to second-line, taxane-based therapy compared with patients who discontinued for toxicity or patient preference (36% vs. 71%; P = 0.01).


Second-line taxane chemotherapy after ixabepilone resulted in a substantial frequency of PSA declines. Although patients with ixabepilone-refractory disease were less likely to respond to second-line taxane chemotherapy, 36% did achieve a PSA response. These findings were consistent with incomplete clinical cross-resistance between the taxanes and the epothilones. Cancer 2006. © 2005 American Cancer Society.

Although androgen deprivation remains the backbone of treatment for patients with advanced prostate carcinoma, ultimately, all patients will develop disease progression despite castrate levels of testosterone. The increasing use of prostate-specific antigen (PSA) testing has resulted in the detection of androgen-independent or hormone-refractory prostate carcinoma (HRPC) earlier in the course of the illness. Secondary hormonal manipulations are effective in some patients, but the duration of response of such agents tends to be short; and, to date, the use of these agents has not been shown to impact survival. The earlier detection of HRPC also has led to earlier use of chemotherapy.

Recently, it was demonstrated that docetaxel chemotherapy prolonged the survival of men with HRPC in 2 large Phase III trials.1, 2 The survival benefit associated with docetaxel represents a 20–25% improvement over mitoxantrone and prednisone therapy (the prior treatment standard). Although those trials have provided the first Phase III evidence of a survival benefit for chemotherapy in men with advanced prostate carcinoma, the median time to progression on docetaxel remains only 6 months. Consequently, there is a growing population of patients who have HRPC that is progressive despite docetaxel-based therapy and for whom there exists no standard of care.

Overall toxicity from taxane-based regimens is relatively modest,1, 2 and patients often remain quite healthy and desire further therapy. The role and utility of second-line chemotherapy in HRPC largely remains undefined. Very few clinical trials have been performed, and the trials that have been reported often have combined untreated and pretreated patients, and they lacked consistency in prior therapy.3–10 Therefore, it is difficult to reach any conclusions about the utility or activity of second-line chemotherapy.

Epothilones are nontaxane tubulin polymerization agents derived from the myxobacterium Sorangium cellulosum.11 The epothilones bypass the multidrug resistance (MDR) protein and tubulin mutation modes of resistance, which play significant roles in taxane resistance.12

Because of their preclinical properties, these agents are undergoing extensive evaluation in multiple tumor types.12 In particular, a Phase I/II multiinstitutional, randomized study of the epothilone-B analog ixabepilone (BMS-247550) with or without estramustine demonstrated substantial activity according to PSA and objective response proportions.13 To begin to address the questions of the role of second-line chemotherapy and the clinical cross-resistance of the taxanes and the epothilones in HRPC, we performed a retrospective analysis of second-line taxane chemotherapy in a cohort of patients who were treated previously with the epothilone analog, ixabepilone.



The records of patients with HRPC from three institutions who were treated on a front-line chemotherapy trial of ixabepilone with or without estramustine were reviewed. Patients were followed for treatment response to subsequent chemotherapy. Patients were eligible for inclusion in this analysis if they received at least two cycles of ixabepilone followed by chemotherapy that contained paclitaxel or docetaxel (single-agent or multiagent). These investigations were approved by the Institutional Review Boards of each institution.


The primary endpoint for this retrospective analysis was the proportion of patients who achieved a posttreatment decline in PSA level ≥ 50%, as defined by the Consensus Criteria for Phase II studies in HRPC.14 The time to PSA progression (defined by the Consensus Criteria) also was evaluated.14 Survival was measured from the time patients received their first dose of second-line taxane chemotherapy to the time of death or the date of last contact. The effect of several pretreatment variables on the second-line PSA response proportion was evaluated. These variables included prior response to ixabepilone-based therapy, discontinuation because of the ixabepilone (toxicity/patient preference vs. disease progression), and the inclusion of estramustine in the front-line, ixabepilone-based therapy.

Statistical Analysis

Descriptive statistics (mean, median, proportion, 95% confidence intervals [95% CI]) were calculated to characterize patient characteristics as well as ixabepilone and taxane-based treatment response outcomes. The Fisher exact test was used to compare response proportions. The time to PSA progression, as defined by the Consensus Criteria, was measured from the start of treatment separately for first-line and second-line chemotherapy.14 Survival was measured from the start of taxane chemotherapy to the time of death or the date of last contact. The time to PSA progression and survival with taxane-based chemotherapy were analyzed using the Kaplan–Meier product-limit method.


Patient Characteristics

Forty-nine patients were identified who received both first-line ixabepilone and second-line taxane-based chemotherapy. The mean age at the start of first-line ixabepilone was 66 years (range, 45–80 yrs). Taxane-based second-line chemotherapy regimens are listed in Table 1. The most common regimen was docetaxel plus estramustine (59% of patients). The PSA response proportion to first-line ixabepilone in this group of patients was 63% (95%CI, 48–7%) (Table 2). The median number of chemotherapy cycles was five cycles of first-line chemotherapy and five cycles of second-line chemotherapy. The median follow-up was 17.6 months for patients who received first-line ixabepilone and 8.1 months for patients who received second-line taxanes. The median time to progression on first-line ixabepilone was 4.6 months, the median time from start of first-line treatment to the start of second-line treatment was 7.0 months, and the median pretreatment PSA level before second-line chemotherapy was 182 ng/mL.

Table 1. Patient Characteristics
CharacteristicNo. of patients (%)
No. of patients who received second-line taxane49
Age in yrs 
Interval between first-line and second-line chemotherapy (mos) 
Gleason score 
 No score available2
First-line chemotherapy 
 Ixabepilone monotherapy21 (43)
 Ixabepilone/estramustine28 (57)
Second-line regimens 
 Docetaxel/estramustine29 (59)
 Docetaxel10 (20)
 Paclitaxel/estramustine/carboplatin 3 (6)
 Docetaxel/carboplatin 3 (6)
 Paclitaxel/estramustine 2 (4)
 Docetaxel/cyclophosphamide 1 (2)
 Docetaxel/calcitriol 1 (2)
Table 2. Results
VariableFirst-line ixabepiloneSecond-line taxane
  1. 95% CI: 95% confidence interval; PSA: prostate-specific antigen.

PSA response  
 No. of patients3125
 Response rate (%)6351
 95% CI (%)48–7736–66
No. of cycles  
Median follow-up (mos)17.68.1
Median time to PSA progression (mos)4.64.6

Clinical Outcomes from Second-Line Chemotherapy

Fifty-one percent of patients who received taxane-based second-line chemotherapy achieved a PSA response (95% CI, 36–66%) (Table 2). For responders, the median nadir PSA was 65 ng/mL (range, 0.8–420.0 ng/mL), and the median percentage PSA decrease was 69% (range, 54–99%). The median time to PSA progression on second-line chemotherapy was 4.6 months. A prior ixabepilone PSA response was not associated with a significantly increased PSA response proportion to second-line taxane (P = 0.06) (Table 3). However, PSA responses in patients who had stopped ixabepilone for toxicity or because of patient preference were significantly higher compared with the PSA responses in patients who stopped ixabepilone for disease progression (P = 0.01) (Table 3). Estramustine used in the first-line setting with ixabepilone did not appear to affect the PSA response proportion observed in second-line taxane combination therapy (Fig. 1). Among the patients who received taxane-based, second-line chemotherapy, 49% of patients remained alive 12 months after the start of second-line chemotherapy, with a median survival of 10.7 months (Fig. 2). The median survival in this cohort of patients from the start of ixabepilone first-line chemotherapy was 22.7 months.

Table 3. Effect of Pretreatment Variables on Prostate-Specific Antigen Response
Pretreatment variableSecond-line response
  1. 95% CI: 95% confidence interval; NS: not significant.

First-line ixabepilone response 
 No. of responders (n = 31 patients)19
  Response rate (%)61
  95% CI (%)42–78
 No. of nonresponders (n = 18 patients)6
  Response rate (%)33
  95% CI (%)13–59
 P value0.06
First-line ixabepilone reason for discontinuation 
 No. who stopped for progression (n = 27 patients)10
  Response rate (%)36
  95% CI (%)19–56
 No. who stopped for toxicity or patient preference (n = 21 patients)15
  Response rate (%)71
  95% CI (%)48–89
 P value0.01
Prior first-time estramustine 
 No. who received prior estramustine (n = 28 patients)14
  Response rate (%)50
  95% CI (%)31–69
 No. who did not receive prior estramustine (n = 21 patients)11
  Response rate (%)52
  95% CI (%)30–74
 P valueNS
Figure 1.

This diagram illustrates patient distribution according to estramustine use. The prostate-specific antigen response proportion for second-line therapy is listed for each subgroup. The impact of first-line estramustine on subsequent second-line response did not appear to be dramatic.

Figure 2.

The median survival from the start of taxane therapy was 10.7 months for patients with hormone-refractory prostate carcinoma, and 48% of patients remained alive at 12 months.


With the recent demonstration of a survival benefit for taxane-based chemotherapy in patients with HRPC, active and tolerable second-line regimens are necessary to provide patients with additional therapeutic options, particularly because the time to disease progression after docetaxel-based chemotherapy remains only 6 months. Ixabepilone chemotherapy for HRPC was associated with high PSA response proportions and objective responses in a recently completed Phase II, front-line chemotherapy trial.13 In this article, we describe the responses of patients with HRPC from that trial who were treated subsequently with second-line taxane chemotherapy. We observed that this group of patients had a second-line, taxane-based chemotherapy PSA response proportion of 51%. In fact, the cohort had a PSA response proportion to second-line taxane therapy similar to that observed with first-line taxane chemotherapy.1, 2, 15 Although it was limited by the retrospective nature of these data, the demonstration of high PSA response proportions with second-line chemotherapy suggests that second-line chemotherapy for HRPC is feasible and that it should be evaluated further in a prospective fashion to determine whether it provides clinical benefit.

Preclinical data on the epothilone analogs demonstrate that this class of drugs has antitumor activity in taxane-resistant carcinoma cell lines and xenograft models. The epothilones bypass the common modes of taxane resistance (MDR over-expression and tubulin mutation). Based on this retrospective analysis of patients with prostate carcinoma who were treated with both an epothilone analog and a taxane, it appears that taxanes and epothilones demonstrate partial clinical cross-resistance in HRPC. However, 36% of patients who experienced disease progression on ixabepilone demonstrated PSA responses to taxane-based second-line chemotherapy. This response proportion is substantially higher than what was observed with second-line chemotherapy in other solid tumors.16–19 These data support the concept that there is not complete clinical cross-resistance between the taxanes and the epothilones and provide the rationale for a subsequent Phase II clinical trial of second-line ixabepilone in patients with taxane-refractory HRPC.

The median survival for patients who received second-line chemotherapy for HRPC in this cohort was 10.7 months. Although this sample clearly was biased by patient selection, the retrospective nature of the data, the absence of patients with very aggressive disease (i.e., patients with disease that progressed rapidly after first-line chemotherapy who died), the absence of patients with less aggressive disease (i.e., patients who have not yet experienced disease progression after ixabepilone or who continue to receive ixabepilone therapy), this estimate may provide some insight for planning Phase III trials of HRPC second-line chemotherapy. Overall, this cohort of patients lived for 22.7 months after the initiation of first-line chemotherapy. Prospective Phase II studies of second-line chemotherapy currently are underway and will provide a more accurate estimate of overall survival for future Phase III studies.

Although the sample size of the current cohort was small, there was no readily apparent impact of the use of first-line estramustine on subsequent second-line PSA responses (50% vs. 52%; nonsignificant). These results are consistent with the observation from the front-line taxane studies suggesting that the addition of estramustine is unlikely to provide a substantial survival benefit in the first-line setting.1, 2 Toxicities seen with the addition of estramustine are considerable, including thromboembolic events, nausea, emesis, and fatigue. Nevertheless, the role of estramustine treatment in the second-line setting remains undefined and could not be addressed by this retrospective review.

A prospective evaluation of second-line chemotherapy clearly is necessary. Mitoxantrone is being used as second-line treatment throughout the oncology community without evidence of its activity in this setting. Given the preclinical properties and the data presented here, ixabepilone and the other epothilones clearly are candidates for further testing as second-line chemotherapy. A Phase II study evaluating ixabepilone as second-line chemotherapy in patients with HRPC is ongoing.