Phase II, Multicenter, Randomized Trial of Docetaxel plus Prednisone with or Without Cediranib in Men with Chemotherapy-Naive Metastatic Castrate-Resistant Prostate Cancer.

LESSONS LEARNED: The negative results are consistent with the negative results of large phase III trials in which docetaxel plus antiangiogenic agents were used in patients with metastatic castrate-resistant prostate cancer (mCRPC).The negative data underscore that, despite a sound biological rationale and supportive early-phase clinical results, adding antiangiogenic agents to docetaxel for mCRPC is a great challenge.
BACKGROUND: Inhibition of vascular endothelial growth factor (VEGF) signaling abrogates tumor-induced angiogenesis to constrain tumor growth, and can be exploited therapeutically by using cediranib, an oral tyrosine kinase inhibitor of VEGF receptor signaling. Our preliminary phase I trial data showed that adding cediranib to docetaxel plus prednisone (DP) was safe and feasible, with early evidence for efficacy in patients with metastatic castrate-resistant prostate cancer (mCRPC).
METHODS: This multicenter phase II trial assessed whether adding cediranib to DP improves efficacy of DP in patients with mCRPC. Chemotherapy-naive patients with mCRPC were randomly assigned to receive either docetaxel (75 mg/m2 intravenously every 3 weeks) with prednisone (5 mg twice daily) plus cediranib (30 mg once daily; the DP+C arm) or DP only (the DP arm). The primary endpoint was to compare 6-month progression-free survival (PFS) rate between the two arms. Secondary endpoints included 6-month overall survival (OS), objective tumor and prostate-specific antigen (PSA) response rates, biomarkers, and adverse events.
RESULTS: The 6-month PFS rate in a total of 58 patients was only numerically higher in the DP+C arm (61%) compared with the DP arm (57%). Similarly, the 6-month OS rate, objective tumor and PSA response rates, and biomarkers were not significantly different between the two arms. Increased baseline levels of interleukin 6 (IL-6), however, were significantly associated with increased risk of progression. Neutropenia was the only grade 4 toxicity (38% in the DP+C arm vs. 18% in the DP arm).
CONCLUSION: Combining cediranib with docetaxel + prednisone failed to demonstrate superior efficacy, compared with docetaxel + prednisone, and added toxicity. Our data do not support pursuing the combination further in patients with mCRPC. © AlphaMed Press; the data published online to support this summary are the property of the authors.

Discussion

We hypothesized that adding the antiangiogenic agent cediranib (AZD2171), an oral tyrosine kinase inhibitor of VEGF receptor signaling, to docetaxel may improve docetaxel's efficacy in mCRPC. We found in a phase I trial that adding cediranib to DP was safe and feasible with early evidence of efficacy in patients with mCRPC (data not shown). As a follow‐up, we conducted a multicenter, randomized, phase II screening study to further evaluate safety and efficacy of the combination. The primary endpoint was to compare 6‐month PFS rate between the two arms. Secondary endpoints included 6‐month OS, objective tumor and PSA response rates, biomarkers, and adverse events.

Because of poor accrual, the study was discontinued prior to full enrollment of 104 patients. The data that were obtained from 58 chemotherapy‐naive patients with mCRPC recruited in nine participating institutions from December 2007 until December 2011 are reported herein. Thirty patients were randomized to the DP+C arm, of whom 29 received treatment. All 28 patients randomized to the DP arm were treated. The median age was 68 years (range, 52–82 years) in the DP+C arm versus 66 years (range, 51–84 years) in the DP arm.

Chemotherapy‐naive patients with mCRPC were randomly assigned to receive either docetaxel (75 mg/m2 intravenously every 3 weeks) with prednisone (5 mg twice daily) plus cediranib (once daily; the DP+C arm) or DP only (the DP arm). Because of the observed toxic effects, the starting dose of cediranib, 30 mg/daily, was reduced to a 20‐mg daily dose.

The 6‐month PFS rate was slightly numerically improved in the DP+C arm compared with the DP arm. This was not statistically significant, however, given the extensive overlap of the arm‐specific 95% confidence intervals (CIs; 61%, 43%–79% for the DP+C arm vs. 57%, 38%‐77% for the DP arm). Similarly, the duration of censored PFS was slightly longer in the DP+C arm, but this was not statistically significant (hazard ratio [HR] for the DP+C arm, 0.91; 95% CI, 0.51–1.62; p = .74; Fig. 1). Adjustment for age, race, and baseline PSA had a negligible effect (HR, 0.94; 95% CI, 0.52–1.70; p = .84). Moreover, median PFS estimates were not significantly different between the two arms, as evidenced by overlapping 95% CIs (8.0 months, 4.2–11.9 months in the DP+C arm and 6.4 months, 4.8–10.2 months in the DP arm). Similarly, 12‐month PFS rates were not significantly different between the two arms.

Figure 1.

Kaplan‐Meier plot for progression‐free survival by treatment arm. Censored (i.e., progression‐free) patients are indicated by vertical marks in each curve.

Abbreviations: DP, docetaxel at 75 mg/m2 intravenously over 1 hour on day one every three weeks plus prednisone at 5 mg orally twice per day on days 1‐21; DP+C, docetaxel at 75 mg/m2 intravenously over 1 hour on day one every three weeks plus prednisone at 5 mg orally twice per day on days 1‐21 plus cediranib at 20 mg orally once daily on days 1‐21.

The 6‐month OS rate and the duration of censored OS were numerically improved but not significantly different in the DP arm compared with the DP+C arm. The partial tumor and complete PSA response rates were only numerically higher in the DP+C arm compared with the DP arm (53% and 35% vs. 33% and 12%, respectively).

There were no significant interactions observed for biomarkers by treatment arm. For IL‐6, however, a 10 pg/mL increase in the baseline level was significantly associated with an 18% increase in the risk of a progression event. The primary grade 4 toxicity neutropenia was observed in 11/29 (38%) patients in the DP+C arm and in 5/28 (18%) patients in the DP arm.

Overall, adding cediranib to docetaxel + prednisone in patients with mCRPC was associated with numerically higher partial tumor and complete PSA response rates, but without any significant improvement in PFS or OS, and with added toxicity. Our data do not support pursuing the combination further in patients with mCRPC.

Trial Information

Advanced cancer/solid tumor only

Metastatic/advanced

None

Phase II

Randomized

Progression‐free survival

Overall survival

Objective tumor responses

Prostate‐specific antigen

Correlative endpoints

Safety

Our data do not support pursuing the combination further in mCRPC.

Drug Information: Cediranib + Docetaxel + Prednisone

Docetaxel/RP 56976

Taxotere

Small molecule

Microtubule‐targeting agent

75 mg/m2

IV

Docetaxel at 75 mg/m2 over 1 hour on day 1 every 3 weeks

Prednisone

DeCortin/Deltra

Small molecule

Corticosteroid

5 mg per flat dose

p.o.

Prednisone at 5 mg twice daily on days 1–21

Cediranib/AZD2171

Recentin

AstraZeneca

Small molecule

VEGFR

30 mg per flat dose

p.o.

Cediranib 30 mg once daily on days 1–21. Reduced to 20 mg once daily by amendment

Drug Information: Docetaxel + Prednisone

Docetaxel/RP 56976

Taxotere

Small molecule

Microtubule‐targeting agent

75 mg/m2

IV

Docetaxel at 75 mg/m2 over 1 hour on day 1 every 3 weeks

Prednisone

DeCortin/Deltra

Small molecule

Corticosteroid

5 mg per flat dose

p.o.

Prednisone at 5 mg twice daily on days 1–21

Patient Characteristics: Cediranib + Docetaxel + Prednisone

29

0

Metastatic castrate‐resistant prostate cancer (stage IV)

Median (range): 68 years (52–82 years)

Median (range): none

0 — 12

1 — 17

2 — 0

3 — 0

Unknown — 0

Complete details of patient characteristics are shown in Table 1.

Table 1.

Baseline and clinical characteristics of patients by treatment arm

DP+C, docetaxel at 75 mg/m2 intravenously over 1 hour on day 1 every 3 weeks + prednisone at 5 mg orally twice daily on days 1–21 + cediranib at 20 mg orally once daily on days 1–21.

DP, docetaxel at 75 mg/m2 intravenously over 1 hour on day 1 every 3 weeks + prednisone at 5 mg orally twice daily on days 1–21.

Race data were missing for one patient in the DP arm.

Abbreviations: C, cediranib; D, docetaxel; ECOG PS, Eastern Cooperative Oncology Group performance status; P, prednisone; PSA, prostate‐specific antigen.

Gleason grade <7, n (%): 2 (7)

Gleason grade = 7, n (%): 10 (37)

Gleason grade >7, n (%): 15 (56)

Unknown Gleason grade, n = 2

Patient Characteristics: Docetaxel + Prednisone

28

0

Metastatic castrate‐resistant prostate cancer (stage IV)

Median (range): 66 years (51–84 years)

Median (range): none

0 — 10

1 — 17

2 — 1

3 — 0

Unknown — 0

Complete details of patient characteristics are shown in Table 1.

Gleason grade <7, n (%): 2 (8)

Gleason grade = 7, n (%): 9 (35)

Gleason grade >7, n (%): 15 (58)

Unknown Gleason grade, n = 2

Primary Assessment Method: Cediranib + Docetaxel + Prednisone

Table 2.

PFS, OS, and objective tumor and PSA responses of patients by treatment arm

DP+C, docetaxel at 75 mg/m2 intravenously over 1 hour on day 1 every 3 weeks + prednisone at 5 mg orally twice daily on days 1–21 + cediranib at 20 mg orally once daily on days 1–21.

DP, docetaxel at 75 mg/m2 intravenously over 1 hour on day 1 every 3 weeks + prednisone at 5 mg orally twice daily on days 1–21.

PFS and OS (6‐month and 12‐month values) are shown as rates in percentages.

Objective tumor responses are shown as number of patients and percentages of patients with measurable disease (15 and 9 on the DP+C and the DP arm, respectively).

PSA responses by PCWG2 criteria are shown as number of patients and percentages of evaluables (26 per each arm).

Abbreviations: C, cediranib; CI, confidence interval; CR, complete response; D, docetaxel; P, prednisone; PD, progressive disease; PCWG2, Prostate Cancer Working Group; PFS, progression‐free survival; PR, partial response; PSA, prostate‐specific antigen; OS, overall survival; SD, stable disease.

Secondary Assessment Method: Cediranib + Docetaxel + Prednisone

Correlative endpoints

Serum markers

See Figure 3 and Table 4

Figure 3.

VEGF levels by treatment arm and time point. Multiple box plot of VEGF levels (in pg/mL) for each treatment arm is shown at all four time points of measurement. The effective sample size (n) is shown above each individual box plot.

Abbreviations: DP, docetaxel at 75 mg/m2 intravenously over 1 hour on day one every three weeks plus prednisone at 5 mg orally twice per day on days 1‐21; DP+C, docetaxel at 75 mg/m2 intravenously over 1 hour on day one every three weeks plus prednisone at 5 mg orally twice per day on days 1‐21 plus cediranib at 20 mg orally once daily on days 1‐21; VEGF, vascular endothelial growth factor.

Table 4.

Hazard ratios for progression by each baseline biomarker

The data for both univariate and bivariate models are shown as hazard ratios (95% CI). Each hazard ratio shows the multiplicative change in the risk of a progression event associated with a 10 pg/mL increase in the level of each biomarker, except for VEGF‐C, for which a 100 pg/mL increase was used.

The univariate model includes only a biomarker.

The biovariate model includes both a biomarker and a treatment arm.

HR significantly different from unity before adjustment for multiple comparisons (p = .01).

Abbreviations: CI, confidence interval; IL, interleukin; VEGF, vascular endothelial growth factor.

Median OS

30

Kaplan‐Meier

Not reached

Objective tumor responses

15

RECIST 1.0

n = 0 (0%)

n = 8 (53%)

n = 6 (40%)

n = 1 (7%)

n = 14

Serum PSA

26

PCWG2 criteria

n = 9 (35%)

n = 7 (27%)

n = 9 (35%)

n = 1 (4%)

n = 3

Primary Assessment Method for Phase II Docetaxel + Prednisone

Secondary Assessment Method for Phase II Docetaxel + Prednisone

Correlative endpoints

Serum markers

See Figure 3 and Table 4

Median OS

28

Kaplan‐Meier

Not reached

Objective tumor responses

9

RECIST 1.0

n = 0 (0%)

n = 3 (33%)

n = 6 (67%)

n = 0 (0%)

n = 19

Serum PSA

26

PCWG2 criteria

n = 3 (12%)

n = 14 (54%)

n = 8 (31%)

n = 1 (4%)

n = 2

Adverse events are shown in Table 3.

Table 3.

Grade 3 and 4 adverse events by treatment arm

DP+C, docetaxel at 75 mg/m2 intravenously over 1 hour on day 1 every 3 weeks + prednisone at 5 mg orally twice daily on days 1–21 + cediranib at 20 mg orally once daily on days 1–21.

DP, docetaxel at 75 mg/m2 intravenously over 1 hour on day 1 every 3 weeks + prednisone at 5 mg orally twice daily on days 1–21.

Abbreviations: C, cediranib; D, docetaxel; P, prednisone.

Assessment, Analysis, and Discussion

Study terminated before completion

Did not fully accrue

Our data do not support pursuing the combination further in mCRPC.

This randomized phase II screening trial included two well‐balanced treatment groups with comparable patient demographics and biological characteristics (Table 1; Fig. 2). No statistical differences, however, were found between the two treatment groups in any of the tested outcome parameters (Figs. 1, 3; Table 2). There were no differences in the extent of disease in either arm. Thus, the data do not support our hypothesis that adding the antiangiogenic cediranib to docetaxel plus prednisone (DP) would improve DP efficacy in metastatic castrate‐resistant prostate cancer (mCRPC). Although the results were disappointing, they are consistent with the negative results of large phase III trials of docetaxel and antiangiogenic combination therapy in patients with mCRPC that have been reported since the initiation of our clinical trial [8].

Figure 2.

Consolidated Standards of Reporting Trials diagram. Overview of screened and randomly assigned patients.

Abbreviations: DP, docetaxel plus prednisone; DP+C, docetaxel plus prednisone plus cediranib.

This clinical trial had to close early because of slow accrual. Accrual was ultimately very challenging because of the rapidly changing treatment landscape, which at the time included emerging oral androgen receptor‐targeted agents in competing clinical trials. Such oral drugs held additional appeal for patients with mCRPC, whose only available option at that time was intravenous chemotherapy. Notably, as reported by Massett et al. [9], the National Cancer Institute recognized competing trials as a reason for slow accrual in phase II studies. They proposed to invest in earlier accrual planning as a better strategy for future trials.

It is doubtful, however, that additional patients would have changed the outcome of our study. In patients with mCRPC, there have been 11 randomized phase III trials of docetaxel combined with GVAX, calcitriol, bevacizumab, aflibercept, dasatinib, atrasentan, zibotentan, lenalidomide, or custirsen, with no difference in overall survival (OS) noted between the treatment arms [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], notwithstanding promising early‐phase data of docetaxel combined with sunitinib [21]. Specifically, phase III trials undertaken in patients with mCRPC using docetaxel combined with the antiangiogenic agents bevacizumab [12] or aflibercept [13] did not meet the primary endpoint of improving OS. Moreover, these combinations were associated with greater toxicity compared with the control arms. Similarly, in a phase III trial in patients with mCRPC, the combination of docetaxel with dasatinib, an SRC inhibitor and an antiangiogenic agent, did not improve OS and added toxicity compared with the active control arm [14]. Another combination tested in the phase III clinical trial setting that worsened not only the toxicity but also the OS in patients with mCRPC was docetaxel plus lenalidomide, an immunomodulatory agent with antiangiogenic properties [18]. Other agents that showed a lack of benefit in mCRPC when combined with docetaxel in the phase III setting include atrasentan, an endothelin receptor antagonist, and custirsen (OGX‐011), a second‐generation antisense oligonucleotide designed to inhibit production of the cytoprotective clusterin that is associated with resistance to chemotherapy [20]. Similarly, the second‐generation taxane cabazitaxel, approved as a second‐line treatment after docetaxel failure, when combined with custirsen did not improve OS in patients with mCRPC [19].

In our trial, the adverse events noted in both arms were as anticipated, with the majority of them likely related to docetaxel (Table 3). Grade 4 neutropenia was higher in the DP+C arm compared with the DP arm. Treatment with cediranib also resulted in considerable fatigue and hypertension that led to permanent dose reduction. As the cediranib dose was reduced, adverse events became somewhat more manageable.

We showed that increased basal blood levels of interleukin 6 (IL‐6; Table 4) were significantly positively associated with PFS, suggesting potential prognostic value of IL‐6. Elevated blood levels of IL‐6 were associated with castration‐resistant prostate cancer or prostate cancer metastatic to bone, and correlated negatively with tumor survival [22], [23], [24]. In addition, blood IL‐6 levels were associated with resistance to chemotherapy in prostate cancer [24], [25]. In our study, however, neither IL‐6 levels nor the levels of any of the other four biomarkers (i.e., vascular endothelial growth factor [VEGF], VEGF‐C, VEGF‐D, and IL‐8) were affected by treatment. It is likely that the short duration of our study and the small number of tested samples were not adequate to identify potential predictive biomarkers of responsiveness to cediranib in combination with docetaxel.

Biomarkers are usually used in selecting patients when administering targeted drugs, and this approach can improve treatment outcome. A distinguishing feature between antiangiogenic drugs and other targeted therapies, however, is that the former are administered to unselected patients within approved indications [26]. There is a need, therefore, to develop predictive biomarkers to identify patients with mCRPC who are more likely to respond to antiangiogenic drugs. As suggested by Kelly et al. [12] the success of future phase III trials in mCRPC depends on identifying critical biomarkers that enrich the study population for the targeted therapy in order to better understand the association between PFS, prostate‐specific antigen response, and objective tumor response as intermediate markers for OS in this patient population.

Overall, our study provides additional evidence supporting the view that docetaxel in combination with any targeted agent has not been successful in patients with mCRPC. The negative data from our study underscore the challenges of adding novel agents to standard treatments for mCRPC, particularly docetaxel, even in the context of a sound biological rationale and supportive early‐phase clinical results. Because promising preclinical data do not often translate to an OS benefit in the clinic, additional preclinical and early‐phase studies for docetaxel combinations are warranted. Heterogeneity of mCRPC is a key factor contributing to poor correlation between preclinical findings and patient data. To meet this challenge, novel targeted agents with complementary mechanisms of action may improve outcomes of docetaxel‐based therapies [27]. The importance of carrying out additional preclinical and early‐phase studies for docetaxel combinations cannot be overemphasized to inform future trial design, such as an ongoing phase III trial (ARASENS; ClinicalTrials.gov identifier NCT02799602) using docetaxel plus darolutamide (ODM‐201), an oral investigational high‐affinity androgen receptor antagonist, in patients with metastatic castration‐sensitive prostate cancer [28].

PFS, OS, and objective tumor and PSA responses of patients by treatment arm are shown in Table 2.

PFS, OS, and objective tumor and PSA responses of patients by treatment arm are shown in Table 2.