A RANDOMIZED TRIAL OF ADJUVANT THALIDOMIDE VERSUS OBSERVATION IN PATIENTS WITH COMPLETELY RESECTED HIGH RISK RENAL CELL CARCINOMA

Abstract

OBJECTIVES

Thalidomide has demonstrated modest activity in patients with metastatic renal cell carcinoma (RCC). We evaluated the role of adjuvant thalidomide after complete resection of high-risk RCC.

METHODS

Eligibility criteria for enrollment on this randomized trial included any histological subtype, T2 (high grade, any N), T3/T4 (any grade, any N), or node positive (any grade, any T) RCC. Eligible patients were randomized to observation or to receive thalidomide 300 mg daily for 24 months. Patients were followed up until disease recurrence or death. The primary objective of this trial was to evaluate the effect of adjuvant thalidomide on recurrence-free survival (RFS) after nephrectomy for high-risk RCC.

RESULTS

After 46 patients were enrolled, the trial was stopped at a median follow up of 43.9 months (range 9.7–74.2). Patients on the thalidomide arm had inferior 2- and 3- year probabilities of recurrence-free survival, compared to controls (47.8% vs. 69.3% and 28.7% vs. 69.3%, respectively; p=0.022). The 2- and 3- year cancer-specific survival was similar for both groups. All observed deaths were attributable to RCC (p=0.392). By multivariate analysis, tumor size and grade predicted recurrence (p= 0.001 and 0.013) and kidney cancer-specific death (p=0.002 and 0.014). Thalidomide treatment, however, was not an independent predictor of recurrence or cancer-specific mortality.

CONCLUSIONS

In this small randomized controlled trial, adjuvant thalidomide therapy after complete resection of high-risk RCC did not improve the 2- and 3-year RFS rates or cancer specific death rates.

In 2007, approximately 210,000 people worldwide were diagnosed with renal cell carcinoma (RCC), the third leading cause of death among genitourinary malignancies and the 12^th leading cause of cancer death overall.¹ Of these patients, roughly 30% presented with metastatic disease at the time of diagnosis, while an additional 20% to 30% with clinically localized disease at the time of surgical therapy developed metastases.² Patients with locally advanced RCC are at a significant risk for progression and death from their renal tumors due to the presence of adverse pathologic features, such as large tumor size, presence of extra-parenchymal tumor extension, associated venous tumor thrombus and loco-regional lymph node involvement. Development of overt metastatic disease in these patients is associated with a median survival duration of 6 to 10 months and a 2-year survival rate of only 10% to 20%.^2,3 These dismal statistics highlight the need for effective adjuvant therapy for properly selected patients with surgically resectable RCC.

Targeted molecular therapies, including vascular endothelial growth factor (VEGF) neutralizing antibodies and tyrosine kinase inhibitors, which block the receptors for VEGF have demonstrated previously unprecedented response rates in patients with metastatic RCC.^4–6 Thalidomide has been shown to have immunomodulatory and anti-angiogenic properties. In animal models, thalidomide has been shown to reduce the expression of pro-angiogenic factors, such as VEGF, basic fibroblast growth factor, and tumor necrosis factor alpha.^7–11 Thalidomide has been shown to be active in several malignant conditions, including Kaposi’s sarcoma, multiple myeloma, and hormone-refractory metastatic prostate cancer.^12–14 In single-arm phase II trials, thalidomide has demonstrated modest activity in metastatic RCC, with a subset of patients experiencing prolonged progression-free survival.^15–18

In view of the reported activity of this agent in metastatic renal cancer, we planned a randomized trial comparing thalidomide versus observation in patients who underwent complete resection of locally advanced RCC.

PATIENTS and METHODS

Patients

The trial was undertaken with the approval and oversight of the Institutional Review Board at the University of Texas M.D. Anderson Cancer Center. Patients were eligible for enrollment, if they had completely resected locally advanced high-risk RCC, as defined by one of the following criteria: pT2 (Fuhrman grade 3 or 4), pT3a-c, T4, or N1–2 disease resected to no evidence of residual disease. RCC anatomic staging was assigned according to the American Joint Committee on Cancer (AJCC) 2001 TNM classification.¹⁹ Tumor grade was determined using the Fuhrman grading system. Tumor histology was classified according to 2001 WHO criteria and all tumor histological subtypes were considered eligible for inclusion. Patients must have recovered from any effects of surgery, which must have been performed within 30 days of enrollment.

Randomization, Treatment, and Follow-Up

At the time of enrollment, patients were randomized in a 1:1 to treatment with thalidomide or observation. All patients had confirmation of pathologic stage, grade, and margin status by the designated study pathologist.

Patients randomized to treatment received thalidomide orally on a daily basis, with an intent to continue therapy for two years or until disease recurrence. The starting dose was 100 mg/day for 2 weeks, then 200 mg/day for 2 weeks, followed by the maximum dose of 300 mg/day. The entire dose was administered in the evening to minimize the effects of sedation. To minimize potential for teratogenecity all patients were enrolled in System for Thalidomide Education and Prescribing Safety (S.T.E.P.S) program (Celgene, Warren, New Jersey).

Toxicity assessments were performed at every clinic visit and graded according to the National Cancer Institute Common Toxicity Criteria (version 2). Thalidomide dose was decreased by 100 mg increments if grade 2 peripheral neuropathy developed. Thalidomide dose was decreased in 50 mg increments at follow-up for other grade ≥2 adverse effects. Thalidomide therapy was discontinued for unacceptable toxicity defined as unpredictable, irreversible, grade 4 non-hematologic toxicity, or grade 3 neuropathy.

Patients were continued on study until disease recurrence, drug toxicity necessitating termination of therapy, refusal of continuing with the protocol requirements, or death.

Objectives and Statistical Analysis

The primary objective of this randomized trial was to evaluate the effect of adjuvant thalidomide on recurrence-free survival (RFS) after complete resection of high-risk locally advanced RCC, compared with observation alone. RFS was calculated from the date of randomization to documented disease recurrence or death. Disease recurrence was defined as local tumor recurrence or distant metastasis confirmed by physical examination, imaging, or both. Management at the time of failure was at investigator and patient option. The study was designed to detect a 50% reduction of relapse rate with a power of 80%, using a two-sided test and a type I error rate of 0.05. For calculation of sample size, a relapse rate for the observation arm was assumed to be 35% at 2 years after surgery. A 50% reduction of relapse rate to 17.5% at 2 years after surgery corresponds to reduction of hazard ratio from 1 to 0.48 for the treatment arm. With these assumptions, the total number of patients to be accrued was 220.

Secondary objectives were cancer-specific survival (CSS) and overall tolerability and safety of thalidomide.

Two interim analyses were planned to allow for the early termination of the trial in light of evidence that the experimental arm is superior or inferior to the observation arm. In order to provide the overall significance level of 0.05 for the study, the interim analyses had a Lan-DeMets monitoring boundary with an O’Brien-Fleming stopping rule.²⁰ The nominal significance levels for the interim and final analyses were based on the number of relapses observed. The interim analyses were planned when 18 and 37 out of the expected 55 recurrences have been observed.

RESULTS

Between December 2001 and December 2005, 46 patients (23 in each study arm) were accrued to this trial. The treatment and observation arms were matched for patient and tumor characteristics (Table 1). Per-protocol interim analysis, performed in January, 2006, when 18 of expected 55 events occurred, revealed a less than 1% chance of observing the minimum 50% reduction in relapse rate if the trend of events continued with full accrual. Consequently it was concluded, that given the minimal likelihood that adjuvant thalidomide would demonstrate the clinically significant benefit projected, continued accrual to the trial was not justified.

Table 1.

Clinical and pathologic characteristics of 46 patients randomized to thalidomide or observation after nephrectomy for high risk RCC.

	All	Thalidomide	Observation	p
No. Patients (%)	46	23 (50.0)	23 (50.0)
Med. follow up (mo.) (range)	43.9 (9.7–74.2)	50.4 (10.8–71.1)	42.2 (9.7–74.2)	0.563
Median Age (yrs) ± SD	58.0±11.6	58.0±12.1	57.0±11.2	0.861
Gender (%)
Female	12 (26.1)	5 (21.7)	7 (30.4)
Male	34 (73.9)	18 (78.3)	16 (69.6)	0.738
ECOG PS (%)
0	35 (69.6)	17 (73.9)	18 (78.3)
1	11 (30.4)	6 (26.1)	5 (21.7)	0.994
T stage (%)
T2	7 (15.2)	2 (8.7)	5 (21.8)
T3a	19 (41.3)	10 (43.5)	9 (39.1)
T3b/c	20 (43.5)	11 (47.8)	9 (39.1)	0.342
N stage (%)
N0/Nx	33 (71.7)	19 (82.6)	14 (60.9)
N1	7 ( 15.2)	2 (8.7)	5 (21.7)
N2	6 (13.1)	2 (8.7)	4 (17.4)	0.258
Fuhrman grade (%)
2	4 (8.7)	3 (13.0)	1 (4.3)
3	23 (50.0)	10 (43.5)	13 (56.5)
4	19 (41.3)	10 (43.5)	9 (39.2)	0.486
LVI (%)
Absent	30 (65.2)	15 (65.2)	15 (65.2)
Present	16 (34.8)	8 (34.8)	8 (34.8)	1.000
Med. tumor size (cm) ± SD	9.8±4.1	10.0±3.5	9.5±4.7	0.974
Histologic subtype (%)
Conventional	34 (73.9)	16 (60.6)	18 (79.3)
Non-conventional	12 (26.1)	7 (39.4)	5 (20.7)	0.738
Sarcomatoid differentiation (%)
Absent	37 (80.4)	18 (78.3)	19 (82.6)
Present	9 (19.6)	5 (21.7)	4 (17.4)	1.000

At a median follow up of 43.9 months (range 9.7–74.2), 18 (78.3%) patients randomized to adjuvant thalidomide and 8 (34.8%) patients randomized to observation had a documented RCC recurrence. Table 2 provides a summary of oncologic outcomes. Briefly, median RFS was 18.5 months in the thalidomide arm, but was not reached in the observation cohort (p=0.022). Likewise, patients treated with thalidomide had inferior 2- and 3- year probabilities of RFS, compared to controls (47.8% vs. 69.3% and 28.7% vs. 69.3%, respectively; p=0.022). Interestingly, in patients who had recurrent disease, median time to disease recurrence was longer in the thalidomide group, compared to the observation group (12.4 months vs. 5.5 months).

Table 2.

Summary of oncologic outcomes in 46 patients randomized to thalidomide or observation after nephrectomy for high risk RCC.

	Thalidomide (N=23)	Observation (N=23)	p
No. with RCC recurrence (%)	18 (78.3)	8 (34.8)
Median RFS — mo (95% CI)	18.5 (0.0–37.8)	Not Reached	0.022
2-year RFS — % (SD)	47.8 (10.4)	69.3 (9.7)
3-year RFS — % (SD)	28.7 (9.7)	69.3 (9.7)
Recurrence site
No. systemic ± local (%)	13 (72.2)	7 (77.7)	0.613
No. local only (%)	5 (27.8)	2 (22.3)	0.787
No. Lung (%)	7 (38.9)	3 (33.3)
No. Liver (%)	4 (22.2)	3 (33.3)
No. Bone (%)	1 (5.6)	1 (11.1)
No. Adrenal (%)	2 (11.1)	2 (22.3)
No. Retroperitoneum (%)	6 (33.3)	4 (44.4)
No. dead from RCC (%)	9 (39.1)	5 (21.7)
Median CSS — mo (95% CI)	71.1 (31.9–110.2)	Not Reached	0.392
2-year CSS — % (SD)	82.6 (7.9)	82.4 (8.0)
3-year CSS — % (SD)	76.7 (9.3)	75.5 (9.9)

The sites of first failure were distant metastases in 72% of the thalidomide-treated patients and 78% of the patients randomized to observation (p=0.613), while regional nodal or isolated local recurrences only were observed in 28% and 22% of study patients, respectively (p=0.787). The pattern of metastatic disease was also similar between the two groups (Table 2).

At the time of this analysis, 9 (39.1%) patients treated with adjuvant thalidomide died of RCC (median CSS 71.1 months), compared to 5 (21.7%) patients randomized to observation (median CSS not reached, p=0.392). The 2- and 3- year CSS was similar in both study arms, and all observed deaths were attributable to RCC (Table 2). In patients who died from RCC, median time to death was longer for the thalidomide-treated patients, compared to that for the patients in the observation group (34.1 months vs. 11.2 months).

The effect of various covariates — study arm, pathologic primary tumor and nodal stage, size and grade — on RFS and CSS was tested with the Cox proportional-hazards model. The small number of patients and events precluded inclusion of all available variables in the multivariate analysis (Table 3). Tumor size and grade each significantly predicted recurrence (p= 0.001 and 0.013) and kidney cancer-specific death (p=0.002 and 0.014), independent of study randomization.

Table 3.

Multivariate Cox regression analysis for the prediction of disease recurrence and kidney cancer-specific mortality.

	Recurrence			Cancer-specific mortality
	Risk ratio	95% CI	p	Risk ratio	95% CI	p
Adjuvant Thalidomide	2.344	0.938 – 5.853	0.068	1.981	0.543 – 7.224	0.301
Pathologic T stage (T2 v. T3)	2.110	0.567 – 7.850	0.265	5.569	0.635 – 47.440	0.121
Pathologic N stage (Nx/N0 v.N1 v. N2)	1.450	0.574 – 3.664	0.432	2.047	0.532 – 7.871	0.297
Fuhrman grade (1 v. 2 v. 3)	2.550	1.221 – 5.323	0.013	4.321	1.337 – 13.965	0.014
Size (1cm increments)	1.196	1.078 – 1.328	0.001	1.229	1.076 – 1.404	0.002

The median duration of therapy administered to 21 patients was 5.9 months (range, 1.0 to 24.4 months). The summary of adverse events (AEs) associated with thalidomide is shown in Table 4. There was no treatment-related mortality. Five grade 3 AEs were experienced by 4 (19.0%) patients treated with thalidomide. The most common AEs were pain, constipation, fatigue, and neuropathy, necessitating at least one dose reduction in 62% of patients. Excluding the patients who experienced RCC relapse before receiving 24 months of adjuvant treatment, only 35.7% (5 of 14) received the planned course of therapy. The majority of the remaining patients discontinued therapy at various time points due to troublesome toxicity.

Table 4.

Thalidomide treatment and adverse effect summary.

No. of patients at randomization No. of patients treated Median time from surgery to initiation of Thalidomide therapy – mo. (range) Median duration of Thalidomide therapy -- mo. (range) Patients with ≥1 dose reduction — no. (%)	23 21 1.3 (0.7–2.1) 5.9 (1.0–24.4) 13 (61.9)
Reason for treatment discontinuation — no. (%)
Completed 24 mo. course of therapy	5 (21.7)
Recurrent disease	9 (39.1)
Adverse event	7 (30.4)
Protocol violation/Patient Request	2 (4.3)
Adverse events – no. (%)	All Grades	Grade 3
Rash	3 (13.0)	0 (0)
Constipation	14 (60.9)	0 (0)
Fatigue/malaise	14 (60.9)	1 (4.3)
Neuropathy	18 (78.3)	1 (4.3)
Deep venous thrombus	1 (4.3)	1 (4.3)
Nausea	4 (17.4)	0 (0)
Pain	23 (100)	1 (4.3)
Edema	11 (47.8)	0 (0)
Dyspnea	12 (52.2)	1 (4.3)
Other	13 (56.5)	0 (0)

DISCUSSION

In this trial, we evaluated the potential role of thalidomide as systemic adjuvant therapy for patients with resected locally advanced RCC. Although patients selected for this trial were at high-risk for RCC recurrence, as documented by 59% relapse rate within reported follow-up, we did not observe the projected improvement in any oncologic outcomes associated with adjuvant thalidomide use, thus resulting in early termination of this clinical trial. Furthermore, we detected inferior RFS, the primary end-point of this trial, in patients treated with adjuvant thalidomide, compared to patients managed with observation. By multivariate analysis, however, adjusting for the effects of pathologic tumor characteristics, adjuvant thalidomide was not an independent predictor of RFS or CSS, suggesting that adverse pathologic features in patients randomized to treatment may have contributed to inferior RFS in the treatment arm of the trial.

Consistent with previous reports, thalidomide therapy had a moderate toxicity profile, with only a minority of patients able to tolerate the planned maximum dose of 300 mg/day and a few patients completing 24 months of adjuvant treatment. The most common side effects observed were constipation, fatigue, pain, and neurological toxicity; however the incidence of grade 3 or 4 AEs was relatively low. Peripheral neuropathy was observed in 78% of patients, although it was mild.

Many other treatment strategies have been evaluated in the adjuvant setting for RCC, but none has translated into improvement of oncologic outcomes.^21–25 The choice of thalidomide as adjuvant treatment was based on the following rationale. At the time of trial design, thalidomide had been shown to eradicate various solid tumors in animal models by inducing apoptosis in tumor neo-vasculature, and demonstrated considerable activity in angiogenic tumors, such as multiple myeloma and Kaposi’s sarcoma.^7,10–13 In addition, emerging phase II data, at that time, had suggested modest anti-tumor effects using thalidomide as single-agent therapy for metastatic, immunotherapy-refractory RCC. A trial conducted at the M.D. Anderson Cancer Center demonstrated a 7% partial response rate, and an additional 31% of patients had stabilization of their disease.¹⁸ In a similar study from Memorial Sloan-Kettering Cancer Center, no partial responses were obtained, but 64% of heavily pre-treated metastatic RCC patients had stabilization of disease.¹⁵ Finally, the convenience of oral administration of thalidomide and its toxicity profile, compared favorably to immunotherapeutic agents used for management of advanced RCC.

Several limitations of this study warrant discussion. The potential impact of unbalanced randomization coupled with the small number of patients may have been contributing factors to the negative findings of this study. A larger proportion of patients randomized to therapy had extrarenal tumor extension, non-clear cell histology and sarcomatoid features. Because of the small number of patients accrued on this trial, it is possible that these differences may have been important, albeit not statistically significant.

We had chosen a moderate dose of thalidomide (300 mg/day) to maximize clinical activity while limiting systemic toxicity. It should be noted that the doses of thalidomide used in the majority of phase II metastatic RCC trials were significantly higher (600 to 1,200 mg/day).^15–18 Nevertheless, due to toxicity, dose reduction was necessary in 61% of our patients, and excluding patients who suffered RCC relapse while on thalidomide, only 36% completed the planned 24 months of therapy; all of these factors could explain lack of improvement in RFS and CSS seen in this study. Alternatively, through it’s immunomodulatory properties, thalidomide may have exerted a deleterious effect on oncologic outcomes of patients with micrometastatic RCC, highlighting the need for rigorous testing of all adjuvant therapeutics in a context of a clinical trial.

CONCLUSIONS

Our data suggest that single-agent thalidomide used in the adjuvant setting after complete resection of high-risk RCC does not improve oncologic outcomes, and therefore, the risk/benefit ratio clearly does not favor the use of thalidomide in this setting.