Radiographic progression by Prostate Cancer Working Group (PCWG)-2 criteria as an intermediate endpoint for drug development in metastatic castration-resistant prostate cancer

Abstract

Objective

To investigate the association of radiographic progression defined by Prostate Cancer Working Group (PCWG)-2 guidelines and overall survival (OS) in men with metastatic castration-resistant prostate cancer (mCRPC).

Patients and Methods

Two trials that used PCWG-2 guidelines to define progression were analysed: a randomized phase II trial (n = 221) comparing first-line docetaxel-prednisone plus AT-101 or placebo, and a phase III trial (n = 873) comparing prednisone plus sunitinib or placebo after docetaxel-based chemotherapy. Cox proportional hazards regression models were used to estimate the association of radiographic progression with OS. Landmark analyses compared progressing patients with those who had not progressed. Sub-analyses compared patients removed from trial for progression vs other reasons.

Results

An increased risk of death was seen for radiographic progression at landmark times from 6 to 12 months with docetaxel-based therapy (hazard ratio [HR] >1.7 at all time-points). An increased risk of death was also seen with post-docetaxel prednisone alone or with sunitinib for progression at landmark times from 2 to 8 months (HR >2.7 at all time-points). Kendall's τ was 0.50 (P < 0.001) in the setting of docetaxel-based therapy and 0.34 (P < 0.001) in the post-docetaxel setting for association between radiographic progression and death amongst patients with both events. Removal from study due to radiographic progression was associated with a significantly lower OS compared with removal for other reasons in both trials. Limitations of a retrospective analysis apply and there was no central radiology review.

Conclusions

Radiographic progression by PCWG-2 criteria was significantly associated with OS in patients with mCRPC receiving first-line docetaxel-based chemotherapy or post-docetaxel therapy. With external validation as a surrogate endpoint in trials showing survival benefits, the use of radiographic progression-free survival may expedite drug development in mCRPC, which has been hampered by the lack of intermediate endpoints.

Introduction

Drug development for metastatic castration-resistant prostate cancer (mCRPC) is plagued by the absence of surrogate endpoints for overall survival (OS). Given the frequent presence of unmeasurable bone metastases and inadequacy of clinical or PSA-based measures of efficacy, composite endpoints and guidelines have been proposed 1-5. The Prostate Cancer Working Group Prostate Cancer Working Group (PCWG)-2 guidelines recommended time to event endpoints 6, 7. Radiographic progression was defined as either Response Evaluation Criteria In Solid Tumors (RECIST) or bone scan progression defined as two or more new lesions at the initial after treatment assessment confirmed by additional lesions ≥6 weeks later, or two or more new lesions at any time beyond the initial post-treatment assessment, which accounts for the potential misclassification of bone progression due to osteoblastic healing.

Previous studies of the association of older criteria (PCWG-1) for radiographic or PSA progression provided some evidence for an association with OS, but this association was weak and concerns over reliability of the assessments and misclassification of progression led to changes in these criteria 3, 5, 7, 8. A correlation between PCWG-2 defined radiographic progression and OS will facilitate its use as an intermediate endpoint in trials. We examined their relationship in two randomised trials that used PCWG-2 criteria to define radiographic progression 9, 10.

Patients and Methods

Both the CS-205 and SUN-1120 trials continued therapy until progression, which was defined by PCWG-2 criteria or death, whichever came first. OS was the primary endpoint in both trials. The CS-205 trial enrolled 221 patients with chemonaive mCRPC. Of these 220 patients were evaluable (one never received treatment), 110 in the placebo+docetaxel-prednisone (DP) arm and 110 in the AT101+DP arm that received chemotherapy every 3 weeks (1 cycle). Patients were stratified by Eastern Cooperative Oncology Group (ECOG)-performance status (PS) and pain. A maximum of 17 cycles was allowed and AT-101 was not continued after discontinuation of DP. Imaging was obtained every 3 cycles or at symptomatic progression. Progression was defined by radiographic progression determined locally at each institution by PCWG-2 guidelines, RECIST 1.0 or symptomatic progression, but not by PSA progression alone 7, 11. Symptomatic progression was defined as a skeletal-related event (SRE) or the combination of new or worsening ECOG-PS by ≥1 and increase in present pain index score by ≥2 points on two consecutive assessments ≥3 weeks apart. Patients who discontinued for reasons other than progression were followed for both progression and OS. Men in both arms of the CS-205 trial were combined for analysis, as there was no significant difference in OS.

The SUN-1120 trial enrolled 873 patients and compared the combination of prednisone with sunitinib (SP) 37.5 mg orally once daily (n = 584) or placebo (PP, n = 289) for men with mCRPC progressing after docetaxel-based chemotherapy 10. Men were stratified by ECOG-PS and progression type (PSA or radiographic). No maximum number of cycles was stipulated and imaging was performed every 2 cycles (8 weeks) or earlier if required. Progression was defined by radiographic criteria as recommended by PCWG-2 and RECIST 1.0 7, 11. Central radiographic review was initially planned but not performed when the trial failed to meet its primary goal of extension of survival. Symptomatic progression was determined by physician discretion and was not recorded, and PSA progression alone was not considered progressive disease. The pain analysis population was a subset of the overall population that had baseline pain defined as worst pain score by modified Brief Pain Inventory scale ≥4 or average score of ≥3 over 7 days before therapy. Patients who discontinued for reasons other than progression were followed for OS but not for progression. Patients were stratified by ECOG-PS and progression type (PSA or radiographic) and randomised 2:1. Men in both arms of the trial were combined for analysis, as there was no significant difference in OS.

Statistical Methods

OS was calculated using the Kaplan–Meier method. Cox proportional hazards regression models were used to estimate the effect of PCWG-2 defined radiographic progression on OS. Treatment received and baseline stratification factors were used as stratum. Initially, a landmark analysis was performed, comparing all patients who had progressed with patients who had not progressed before the landmark time. Progression was examined as defined by PCWG-2 guidelines and death was excluded. Patients who died or who came off-study for reasons other than radiologic progression before the landmark time were excluded. This analysis was performed using landmark times of 2, 4, 6, 8, 10 and 12 months after randomisation and OS was calculated from the landmark time. Association between time to radiographic progression and OS was evaluated using Kendall's τ statistic using only patients who were observed to have both a progression and death.

A sub-analysis examined the association of the type of progression with OS. For this analysis, OS counted from the day after progression, thereby excluding patients who had progression due to death. Given small numbers, no formal statistical test was performed using the CS-205 dataset. A second analysis was performed comparing OS from the date of radiographic progression between patients who came off-study due to progression with those who came off-study due to reasons other than progression. The landmark analysis described above was repeated with any type of progression (radiological, symptomatic, or skeletal). Statistical significance was defined at α = 0.05 level, and tests and CIs were two-sided.

Results

Patient Characteristics

The primary results of the CS-205 trial have been reported elsewhere 9. Briefly, among a total of 220 evaluable patients, there were 108 patient deaths and 153 patients had progressed after a median (maximum) follow-up amongst survivors of 18.0 (28.8) months. The treatment arms were balanced and outcomes similar for median OS (18.1 vs 17.8 months) and median progression-free survival (PFS; 11.0 vs 10.3 months) for AT-101-DP and placebo-DP, respectively. Reasons for therapy discontinuation included disease progression (23 and 38 patients), death (none and one), subject/sponsor/investigator request (36 and 31) and adverse events (30 and 18). Progression was by bone scan (six and seven patients), SRE (two and three), soft tissue (11 and 24) or symptomatic progression (both four). The number of patients who died without recorded radiographic progression was 29 and 40, respectively.

The SUN-1120 trial accrued 873 men randomised to receive SP (584 patients) or PP (289), and results have been reported 10. Briefly, the median OS was similar for SP and PP (13.1 vs 11.8 months) but the median PFS improved with SP at 5.6 vs PP 4.1 months (P < 0.001 by stratified log-rank). The median (range) treatment duration was 3.2 (0−25.8) and 3.2 (0.2−21.7) months in the SP and PP arms, respectively. Reasons for discontinuation included disease progression (43.9% vs 60.0%), adverse events (27.4% vs 7.4%), death (7.7% vs 4.9%) and consent withdrawal (8.3% vs 8.4%). Of those who progressed, 44%/46% of SP/PP had disease progression observed on bone scan, 16%/15% by RECIST 1.0, 7%/8% on bone scan and RECIST 1.0, and 32%/31% due to neither.

Association of Radiographic Progression at Landmark Times with OS

In the CS-205 trial, 142 patients remained on-study at the 6-month landmark date and they had a median OS estimate of 14.4 (95% CI 11.0–19.5) months (Table 1A). There were 21 patients who had radiographic progression before 6 months, and the median OS for these patients was 5.4 (95% CI 2.8–6.67) months. The hazard ratio (HR) for those who had progressed was 4.42 (95% CI 2.57–7.60, P < 0.001). Regardless of the landmark from 6 to 12 months, there was a statistically significant increased hazard (HR >1.70 at all time-points) of dying for patients with radiographic progression (Fig. 1A,B). At the 12-month landmark, the concordance statistic was 0.567. Kendall's τ was 0.50 (P < 0.001), which indicates a modest association between time to radiographic progression and death, amongst patients with both a recorded radiographic progression and death (39 patients). A sub-analysis with progression defined as any type of progression (radiographic or symptomatic) gave similar results (data not shown).

Table 1. Landmark analyses of association of progression with OS in CS-205 trial* (A) and in SUN-1120 trial^† (B)

Landmark time, months	HR (95% CI)	P	Statistic	No progression		Progression
				N	OS (95% CI)	N	OS (95% CI)
A. CS-205 trial^‡
6	4.42 (2.57–7.60)	<0.001	Median	142	14.4 (11.0–19.5)	21	5.4 (2.8–6.6)
			6-month		84.9 (77.8–89.9)		40.5 (19.6–60.5)
			1-year		54.0 (44.7–62.4)		10.1 (1.7–27.5)
8	3.49 (2.04–5.97)	<0.001	Median	119	15.5 (9.9–NR)	26	4.4 (3.2–6.8)
			6-month		80.5 (72.1–86.6)		40.0 (21.3–58.1)
			1-year		56.5 (46.4–65.5)		6.3 (0.4–24.3)
10	2.43 (1.36–4.35)	0.003	Median	95	14.5 (10.4–NR)	28	4.9 (2.6–8.3)
			6-month		73.9 (63.6–81.7)		44.7 (25.6–62.2)
			1-year		57.9 (46.0–68.1)		22.7 (7.8–42.2)
12	1.70 (0.90–3.21)	0.10	Median	69	12.5 (9.4–NR)	32	6.3 (3.0–13.5)
			6-month		66.5 (53.5–76.6)		52.4 (31.8–69.5)
			1-year		52.1 (35.1–66.6)		36.3 (17.1–55.9)
B. SUN-1120 trial^§
2	2.71 (2.10–3.49)	<0.001	Median	485	15.7 (11.7–15.3)	129	7.4 (5.9–8.5)
			6-month		83.1 (79.5–86.2)		56.0 (46.9–64.2)
			1-year		58.7 (53.9–63.2)		26.7 (18.8–35.3)
4	3.51 (2.69–4.59)	<0.001	Median	270	18.7 (15.5–20.0)	202	6.5 (5.8–7.1)
			6-month		90.2 (85.5–92.9)		53.4 (46.1–60.2)
			1-year		71.7 (65.5–77.0)		25.4 (19.0–32.3)
6	3.78 (2.66–5.37)	<0.001	Median	151	NR	207	7.6 (5.6–9.1)
			6-month		91.0 (85.0–94.7)		56.5 (49.1–63.1)
			1-year		76.1 (67.6–82.7)		34.9 (27.6–42.2)
8	4.53 (2.78–7.38)	<0.001	Median	87	NR	200	7.9 (6.3–10.2)
			6-month		93.0 (85.1–96.8)		58.7 (51.1–65.6)
			1-year		80.1 (68.9–87.6)		30.3 (22.3–38.8)

• *Includes off-study progression in patients removed from trial for reasons other than progression (since trial recorded progression in such patients); ^‡stratified by treatment group and trial baseline stratum of ECOG-PS and pain. ^†Does not include off-study progression in patients removed from trial for reasons other than progression (since trial did not record progression in such patients); ^§stratified by treatment group and trial baseline stratum of ECOG-PS and progression type.

Figure 1

(A) OS and PFS for patients with both radiographic progression and death events in CS-205 trial. (B) OS based on radiographic progression status by 12 months in CS-205 trial.

In the SUN-1120 trial, 485 patients remained on-study at the 2-month landmark date and these patients had a median OS estimate of 15.7 (95% CI 13.7–17.3) months (Table 1B). There were 129 patients who had radiographically progressed before 2 months, and median OS for these patients was 7.4 (95% CI 5.9–8.5) months. The HR for those with radiographic progression before 2 months was 2.71 (95% CI 2.10–3.49) compared with those still on-treatment, which was statistically significant (P < 0.001). Regardless of the landmark from 2 to 8 months, there was a statistically significant increased hazard of dying (HR 2.71–4.53) for patients who had radiographically progressed. Amongst 212 patients who had both a recorded radiographic progression and death, the Kendall's τ was 0.34 (P < 0.001). A sub-analysis including radiographic and symptomatic progression as progression events showed similar results (data not shown).

Association of Type of Progression and OS

In the CS-205 trial, 16 patients had symptomatic progression, and these patients had the worst after progression median survival (5.6 months; Table 2). Those with progression due to measurable disease (53 patients), with a bone scan progression (23) and those with a SRE (eight) had median after progression survival of 7.6, 9.3 and 10.3 months respectively.

Table 2. OS by type of progression in the CS-205 trial

Statistic	Bone scan progression		SRE		Progression of soft tissue		Symptomatic
	N	OS (95% CI)	N	OS (95% CI)	N	OS (95% CI)	N	OS (95% CI)
Median	23	9.3 (5.7–11.0)	8	10.3 (1.1–NR)	53	7.6 (4.7–10.4)	16	5.6 (1.6–6.7)
6-month		67.8 (41.6–84.1)		62.5 (22.9–86.1)		60.8 (44.0–73.9)		35.0 (11.8–59.8)
1-year		18.0 (3.0–43.0)		33.3 (5.6–65.8)		27.8 (12.6–45.3)		0

In the SUN-1120 trial, patients who had progression due to measurable disease only (75 patients) had the best survival after progression (median 12.4 months), while those patients who had bone only progression (217) had the worst survival (median 7.7 months; Table 3). The type of progression was statistically significant (P = 0.004) as a prognostic factor for OS.

Table 3. OS by type of progression in the SUN-1120 trial

Statistic	Measurable only		Bone only		Measurable (± bone)		Bone (± measurable)		Measurable and bone		Clinical
	N	OS (95% CI)	N	OS (95% CI)	N	OS (95% CI)	N	OS (95% CI)	N	OS (95% CI)	N	OS (95% CI)
Median	75	12.4 (8.8–16.4)	217	7.7 (6.4–9.4)	110	12.3 (8.8–15.6)	252	8.2 (6.9–9.4)	35	9.1 (6.4–16.2)	95	9.8 (7.4–12.9)
6-month		77.3 (65.6–85.5)		60.2 (53.3–66.5)		75.3 (65.9–82.5)		61.8 (55.4–67.5)		71.0 (52.8–83.2)		74.4 (64.1–82.2)
1-year		54.3 (40.8–65.9)		31.6 (24.9–38.5)		50.5 (39.7–60.4)		33.2 (26.9–39.7)		42.7 (24.8–59.5)		41.6 (28.8–53.0)

Removal from Trial for Radiographic Progression Versus Other Reasons

In the CS-205 trial, 175 patients were removed from treatment before completion of treatment for reasons other than death. Of these patients, 52 came off-study due to radiographic progression and they had a median OS of 7.8 (95% CI 4.9–10.3) months after coming off-study compared with 10.3 (95% CI 7.4–15.2) months for those who came off for reasons other than radiographic progression (P = 0.053; Table 4).

Table 4. OS by removal for progression or other reasons*

Trial	HR (95% CI)^†	P^†	Statistic	No progression		Radiographic progression		Non-radiographic progression
				N	OS (95% CI)	N	OS (95% CI)	N	OS (95% CI)
CS-205^‡	1.58 (0.99–2.51)	0.053	Median	110	10.3 (7.4–15.2)	52	7.8 (4.9–10.3)	13	6.0 (2.1–NR)
			6-month		72.6 (62.8–80.2)		62.8 (46.0–75.8)		50.0 (20.8–73.6)
			1-year		46.0 (35.7–55.8)		24.0 (10.4–40.6)		25.0 (6.0–50.5)
SUN-1120	1.25 (1.02–1.52)	0.028	Median	391	10.2 (9.6–10.8)	327	8.8 (7.7–10.3)	95	9.8 (7.4–12.9)
			6-month		70.3 (65.2–74.8)		65.3 (59.7–70.3)		74.4 (64.1–82.2)
			1-year		39.9 (33.6–46.1)		37.9 (32.1–43.7)		41.6 (29.8–53.0)

• ^‡Note: patients who completed study treatment and had no observed progression are excluded from this analysis; ^†for comparison of radiographic progression with no progression; *excludes patients who died within 0.5 months of coming off-study.

In the SUN-1120 trial, 327 patients came off-study due to radiographic progression and they had a median OS of 8.8 (95% CI 7.7–10.3) months after coming off-study. In comparison, 391 patients came off-study due to reasons other than radiographic progression, and they had a median OS of 10.2 (95% CI 9.6–10.8) months after coming off-study (P = 0.028; Fig. 2A,B).

Figure 2

(A) OS and PFS for patients with both radiographic progression and death events in SUN-1120 trial. (B) OS based on radiographic progression status by 6 months in SUN-1120 trial.

Discussion

In this post hoc analysis of two randomised trials, radiographic progression by PCWG-2 criteria was significantly associated with poor OS in patients with mCRPC receiving docetaxel-based chemotherapy or post-docetaxel therapy. Shorter time to progression was associated with shorter survival regardless of whether death and symptomatic progression were considered progression events. Removal from trial for radiographic progression was associated with poor OS compared with removal for reasons other than death. Another presentation demonstrated the association of radiographic PFS (by PCWG-2 criteria) and OS in patients with mCRPC receiving pre-docetaxel prednisone alone or with abiraterone in a phase III trial 12. In that trial, radiographic PFS included death as a progression event, but did not include symptomatic progression. An older study reported an association of PFS (including progression or death) with OS; however, progression by PSA criteria was included in the definition of progression, and PCWG-2 criteria for bone scan progression were not used 5. Indeed, in the pre-PCWG2 era, extension of PFS did not translate to extension of OS with second-line satraplatin chemotherapy 13.

In the context of chemotherapy, a PSA level decline of ≥30% within 3 months is moderately associated with OS in the first-line chemotherapy setting 1, 2. However, PSA level declines were not found to be surrogates for survival in the second-line chemotherapy setting 14. PSA progression at 3 months was associated with poor OS 3, 15. Measurable disease response was associated with OS in the setting of chemotherapy, but a minority of patients have measurable tumours 4. The bone scan index, a quantitative measure of tumour burden on bone scintigraphy, and biochemical bone turnover markers (e.g. urinary N-telopeptide) have appeared promising 16, 17. Although circulating tumour cells changes carry prognostic value, all patients do not have detectable circulating tumour cells and costs may be a barrier 18, 19. Positron-emission tomography (PET) scan using ¹⁸F-sodium fluoride and ¹¹C-choline are emerging to identify metastatic disease and monitor tumour burden 20. However, all of these variables were unavailable in these two trials for analysis.

The present study suffers from the limitations of a retrospective analysis and the lack of central radiology review. Both randomised trials did not show superiority for an arm, and the docetaxel-based treatment trial was a phase II trial, while the post-docetaxel trial was phase III. Although both trials used PCWG-2 guidelines and did not deem PSA progression alone to constitute progression, there were other minor differences in the definition of progression. The CS-205 trial used radiographic and symptomatic criteria, while the SUN-1120 trial used radiographic criteria only to define progression and did not objectively capture pain progression. Despite these differences, we isolated the association of radiographic progression with OS. PCWG-2 criteria are essentially arbitrary, and bone scan interpretations may be subject to inter-observer variability. While the present analysis cannot demonstrate surrogacy of radiographic or composite PFS definitions for survival due to the negative survival results of each trial, their relatively strong associations with OS suggests that this endpoint is more strongly associated with mortality than prior definitions. The sub-analyses examining the association of the specific type of progression with OS were under-powered. Potentially, the strength of the association differs based on baseline prognostic factors described in nomograms in the docetaxel-based setting and post-docetaxel settings, but our objective was to show a broad association regardless of risk groups 21, 22. Finally, the magnitude of extension in radiographic PFS that may translate to an extension of OS is unclear, and may vary between different agents. External validation of the present data is warranted in trials showing survival benefits.

Acknowledgments

Ascenta Therapeutics and Pfizer, Inc. for providing the data for analysis.

Conflict of Interest

• G.S.: Research support from Pfizer

• A.J.A.: Research support, speaker for Pfizer

• M.D.G.: Research support from Pfizer

• S.-L.W.: Employment by Pfizer

• J.P.: Employment by Pfizer

• I.C.: Employment by Pfizer

• L.L.: Employment by Ascenta Therapeutics

• B.A.W.: Employment by Ascenta Therapeutics

• E.C.-M.: Employment by Pfizer

• M.L: Employment by Pfizer

• M.R.S: Research support from Pfizer

• M.D.M: Research support from Pfizer

• G.P. and D.M. None

Funding Sources

No funding was provided.

Abbreviations

DP

docetaxel-prednisone

ECOG(-PS)

Eastern Cooperative Oncology Group (performance status)

HR

hazard ratio

mCRPC

metastatic castration-resistant prostate cancer

OS

overall survival

PCWG

Prostate Cancer Working Group Prostate Cancer Working Group

PFS

progression-free survival

PP

prednisone with placebo

RECIST

Response Evaluation Criteria In Solid Tumors

SP

prednisone with sunitinib

SRE

skeletal-related event