Results from a randomized phase II trial of sunitinib and gemcitabine or sunitinib in advanced renal cell carcinoma with sarcomatoid features: ECOG-ACRIN E1808

Abstract

Introduction:

Sarcomatoid renal cancer (sRCC) patients have poor outcomes. EA1808 evaluated sunitinib and gemcitabine (SG) and sunitinib alone (S) in sRCC in a randomized cooperative group phase II trial (NCT01164228).

Patients and Methods:

Pts were aggregated 1:1 to SG (45 pts) or S (40 pts) using a two-stage design. sRCC pts with ≤ 1 prior non-vascular endothelial growth factor tyrosine kinase inhibitor were stratified into prognostic groups: good (clear cell, < 20% sarcomatoid, PS 0), intermediate (20-50% sarcomatoid, PS 0), and poor (non-clear cell or > 50% sarcomatoid or PS 1). The primary endpoint was response rate (RR). For SG, the null RR was 15% and a 30% RR was of interest. For S, a 20% RR was of interest vs. a 5% null rate. Secondary endpoints were progression-free survival, overall survival, and safety.

Results:

Both arms met protocol criteria for stage 2 of accrual. 47 pts were randomized to SG and 40 to S. The SG arm had 9 of 45 evaluable patient responses (RR of 20%; CI = [13% - 31%]) not meeting the predetermined threshold for success. The sunitinib arm met its endpoint with 6/37 (RR of 16%; CI = [9% - 27%]) evaluable responses. Grade ≥ 3 events were experienced by 36 in the SG arm and 17 in the sunitinib arm.

Conclusions:

EA1808 was the largest and first randomized cytotoxic trial for sarcomatoid RCC. Sunitinib alone but not the SG met the preset threshold of success. Cytotoxic chemotherapy is only useful in limited clinical scenarios for sRCC.

Microabstract

Patients with sarcomatoid renal cancer in good, intermediate and poor prognostic groups were randomized 1:1 sunitinib and gemcitabine (SG) versus sunitinib and assessed for response rate in this cooperative group phase II trial. Sunitinib but not the SG met the preset threshold of success Grade ≥ 3 events were seen. EA1808 was the first randomized cytotoxic trial for sarcomatoid RCC.

Introduction

Multiple advances have been made in the treatment of renal cell carcinoma. Advances with VEGF-based inhibitors as well as immunotherapeutics and combinations of those agents have contributed to improved outcomes with regards to overall survival^1-4. However, close to 14,000 patient were estimated to die in 2021 from renal tumors⁵. Moreover, many patients with adverse prognostic features as well as non-clear cell histologies have very poor outcomes ^6-8. Patients who have renal cell carcinoma of any histology with sarcomatoid features have historically worse outcomes and responses to therapy than patients with non-sarcomatoid features ^9-12. The extent of sarcomatoid features also plays a role. Patients with greater than 20% sarcomatoid features have lower response rates to certain systemic treatment modalities.¹⁰ Prior studies have examined ways in which to exert rapid control over particularly aggressive cell types with sarcomatoid features^13,14. Early case reports reported cytotoxic chemotherapy and BCG based combinations in sarcomatoid RCC with notable periods of stable disease.¹⁵ Additional reports looked at the combination of inhibitory vascular endothelial growth factor (VEGFi) based therapies and interferon.¹⁶ Multiple other studies and retrospective analyses examined other forms of systemic agents both alone and together with chemotherapy in sarcomatoid RCC^17-19. Doxorubicin and ifosfamide was tested in a French series of 25 patients, with no objective responses.²⁰ An earlier phase II study ECOG-ACRIN E8802, examined combination doxorubicin and gemcitabine in sarcomatoid renal cell carcinoma patients.²¹ This study showed a 15.4% response rate with those 2 agents. Additional studies have also examined this particular combination chemotherapy in patients with sarcomatoid histologic features and high-grade renal cell carcionoma²². With the introduction of VEGFi anti-angiogenic agents as first line therapy for renal cell carcinoma, a phase II study examining the combination of sunitinib and gemcitabine was examined in patients not only with sarcomatoid features, but also in patients with poor-risk metastatic clear cell renal cell carcinoma. ²³ There was an overall response rate of 26% in this population with benefit appearing more prominently in patients with higher percentage of sarcomatoid cells. ²³ Given the promising results from the earlier combinations of VEGFi therapies and our prior experience with E8802 cytotoxic therapy, plus the lack of standardized therapies for the patients with aggressive sarcomatoid histologies, we performed a follow up phase II randomized prospective trial to compare combination of sunitinib/gemcitabine (S plus G) versus standard of care therapy at the time, sunitinib (S). The primary objective endpoint of this study ECOG-ACRIN E1808, was response rate. Secondary endpoints included progression free survival, overall survival, and toxicity.

Patients and Methods

The trial was reviewed and approved by the ECOG-ACRIN Cancer Research Group (ECOG-ACRIN) Genitourinary committee and the Central Institutional Review Board (CIRB)of the National Cancer Institute. Eligible patients had measurable advanced disease with confirmed component of sarcomatoid histology by central review and ECOG performance status 0 - 2. Patients with collecting duct or medullary carcinoma, as well as patients who received prior systemic cytotoxic chemotherapy were excluded from registration. Additional eligibility criteria are available within the study protocol available via supplementary materials.

Patients with confirmed sarcomatoid RCC were randomized to sunitinib plus gemcitabine (Arm A) or sunitinib alone (Arm B) (Figure 1 Supplemental). Arm A consisted of gemcitabine 1000 mg/m² IV over 30 minutes on days 1, 8, 22, and 29 of each cycle. Sunitinib 37.5 mg was given once daily to Arm A patients on days 1 through 14 and on days 22 through 35 of each cycle. Cycles were 42 days in duration and were be repeated for total of 1 year. Arm B consisted of sunitinib 50 mg daily on days 1 through 14 and days 22 through 35 of each cycle. Each cycle was 42 days and was to be repeated for a duration of one year.

The trial was designed with a two-stage approach. Due to slow accrual, target accrual was reduced from 100 patients to 85 patients. For Arm A (S+G), the combined benefit of gemcitabine and sunitinib would be of interest if the response rate was 30%, while a rate of 15% would suggest ineffectiveness. Assessment of responses took place after 23 eligible and treated patients enrolled in the first stage. If ≥ 4 responses were observed, the arm would remain open for accrual of 20 additional eligible patients. The regimen would be declared a success if overall, 10 or more responses were observed among the first 43 patients in this arm. For Arm B, a response rate of approximately 20% to sunitinib alone was of interest. Further testing would not be pursued if response rate was 5% or lower. Twelve eligible patients were to be accrued initially. If 1 or more responses were observed, there was a plan for accrual of 25 additional eligible patients. Overall, 4 or more responses out of the first 37 eligible patients would be considered a success in this single agent arm.

Response evaluation criteria and solid tumors (RECIST) version 1.1 was used to evaluate response and progression. Patients were evaluated for response every 12 weeks. Complete response (CR) and partial response (PR) were considered positive responses counting towards the predetermined endpoints. Progression free survival (PFS) was identified as the time from randomization to documented progression of disease by RECIST criteria or definite cause or death. Patients alive without progression were censored at the date of last disease assessment. A median PFS of 5.1 months (50% improvement over the median PFS of 3.4 months from E8802 study) was considered significant enough to proceed for further study. Overall survival was defined as time from randomization to death. Patients that were alive as of analysis time were censored at time of last contact. Kaplan Meier survival estimates were utilized.

Results

The study opened on June 17, 2010 and completed November 12, 2015. Data were retrieved for analysis on March 30, 2020 to allow for complete outcome data. The 1:1 randomization, stratified by risk level, utilizing a risk criteria devised prior to the use of the International Metastatic RCC Database Consortium ²⁴ criteria, consisting of good risk (clear cell and < 20% sarcomatoid and PS 0); intermediate risk (20-50% sarcomatoid and PS 0); and poor risk (non-clear cell or greater than 50% sarcomatoid or PS 1). Patients with performance status 2 were eligible and were classified as poor risk.

A total of eighty-seven patients were enrolled, 47 randomized to sunitinib plus gemcitabine and 40 to sunitinib alone. Out of 47 patients enrolled and randomized to sunitinib plus gemcitabine, 1 patient did not receive allocated intervention, and 1 patient was later found to be ineligible, leaving a total of 45 patients ultimately analyzed for efficacy. Forty patients were enrolled and randomized to sunitinib alone, but 1 patient had a progression before starting treatment, 1 patient withdrew consent, and 1 patient was found to be ineligible. (Table 1 Supplemental– case status by treatment arm). Therefore, thirty-seven patients were analyzed for efficacy in this arm (Figure 2 Supplemental, CONSORT Diagram).

The two arms were well balanced with arms having approximately 67% and 65% male and 33% and 35% female, for Arm A and Arm B respectively. Median age was 59 years in the S+G arm and 61 years old in S arm. The majority of patients were Caucasians, comprising of 93% in S+G arm and 89% in S arm. Median percent sarcomatoid histology was 38% in S+G arm vs. 40% in S arm. Patients with clear cell histology made up 70% of both arms. The distribution of sarcomatoid histology was similar between the 2 arms, with more than a third of patients’ specimens having more than 50% sarcomatoid histology in S+G and S (37% vs. 36%) (Table 1). However, a higher proportion of patients with poor performance status (ECOG 2) were enrolled in the combination S+G arm (Supplemental Table 2).

Table 1 –

Patient demographics and disease characteristics

	Sunitinib +gemcitabine (Arm A) n = 45	Sunitinib (Arm B) n = 37	Total n = 82
	N (%)	N (%)	N (%)
Sex
Male	30 (67)	24 (65)	54 (66)
Female	15 (33)	13 (35)	28 (34)
Age, years - median (range)	59 (36 - 79)	61 (44 - 80)	61 (36 - 80)
Race
Asian	0	1 (3)	1 (1)
Black	3 (7)	3 (8)	6 (7)
White	42 (93)	32 (89)	74 (91)
Unknown/Missing	0	1	1
Ethnicity
Hispanic/Latino	3 (7)	0	3 (4)
Non-Hispanic	40 (93)	36 (100)	76 (96)
Unknown/missing	2	1	3
PS
0	20 (44)	20 (54)	40 (49)
1	19 (42)	15 (41)	34 (41)
2	6 (13)	2 (5)	8 (10)
Clinical T-stage
1	6 (14)	6 (17)	12 (15)
2	3 (7)	7 (19)	10 (13)
3	22 (51)	20 (56)	42 (53)
4	9 (21)	2 (6)	11 (14)
T	1 (2)	0	1 (1)
X	2 (5)	1 (3)	3 (4)
Missing	2	1	3
Clinical N-stage
0	10 (24)	16 (46)	26 (34)
1	20 (49)	11 (31)	31 (41)
2	1 (2)	0	1 (1)
D	1 (2)	0	1 (1)
X	9 (22)	8 (23)	17 (22)
Missing	4	2	6
Clinical M-stage
0	12 (32)	9 (29)	21 (30)
1	26 (68)	22 (71)	48 (70)
Missing	7	6	13
Histology
Clear cell	30 (70)	26 (70)	56 (70)
Indeterminate	1 (2)	0	1 (1)
Mixed	2 (5)	1 (3)	3 (4)
Papillary	1 (2)	3 (8)	4 (5)
Unclassified	3 (7)	3 (8)	6 (8)
Other	6 (14)	4 (11)	10 (12)
Missing	2	0	2
% sarcomatoid
Median (range)	38 (1 - 100)	40 (5 - 100)	40 (1 - 100)
Missing	7	4	11
% sarcomatoid
<20%	8 (21)	10 (30)	18 (25)
20-50%	16 (42)	11 (33)	27 (38)
≥50%	14 (37)	12 (36)	26 (37)
Missing	7	4	11

The median number of cycles received in each arm were 3 cycles in Arm A (S+G), versus 2 cycles in Arm B (S). The majority of patients in both arms discontinued treatment due to progression (24/45 patients for sunitinib and gemcitabine vs. 25/37 for sunitinib alone).

Toxicity

All patients that started treatment were included in the toxicity analysis (46 patients in the S+G arm and 37 patients in the S arm). Adverse events led to discontinuation in 13 out of 45 eligible patients in the S+G arm, and 7 out of 37 eligible patients in S arm (Table 2 - Supplemental). For the S+G arm, the most common adverse events included 14 cases of grade 3 & 5 cases of grade 4 neutropenia. There was also grade 3 anemia in 13 patients and 1 case of grade 4 anemia. Grade 3 fatigue was noted in 10 out of 46 patients that started treatment. Grade 4 thrombocytopenia occurred in 6 patients. In the S arm alone, Grade 3 hypertension was noted in 8 out of 37 patients. Additional adverse events in S arm included Grade 3 fatigue in 4 patients and dehydration in 3 patients. Overall, cytopenias were more prevalent in the S+G arm, and there was one Grade 5 adverse event – multi organ failure that occurred during Cycle 1 in S+G arm (Table 2).

Table 2 –

Treatment-related toxicities

Toxicity Type	Treatment Arm
	A Sunitinib +gemcitabine (n=46)			B Sunitinib (n=37)
	Grade			Grade
	3	4	5	3	4	5
	(n)	(n)	(n)	(n)	(n)	(n)
Anemia	13	1	-	-	-	-
Chest pain - cardiac	1	-	-	-	-	-
Heart failure	2	-	-	-	-	-
Myocardial infarction	1	-	-	-	-	-
Pericardial effusion	1	-	-	1	-	-
Fatigue	10	-	-	4	-	-
Fever	2	-	-	-	-	-
Malaise	1	-	-	-	-	-
Multi-organ failure	-	-	1	-	-	-
Pain	-	-	-	1	-	-
Palmar-plantar erythrodysesthesia	1	-	-	1	-	-
Rash maculo-papular	-	-	-	1	-	-
Skin ulceration	-	-	-	1	-	-
Diarrhea	1	-	-	1	-	-
Nausea	1	-	-	1	-	-
Stomach pain	-	-	-	1	-	-
Vomiting	1	-	-	1	-	-
Sepsis	-	-	-	-	1	-
Fall	1	-	-	-	-	-
Alanine aminotransferase increased	1	-	-	-	-	-
Aspartate aminotransferase increased	1	-	-	-	-	-
Blood bilirubin increased	2	-	-	-	-	-
Creatinine increased	1	-	-	-	-	-
GGT increased	1	-	-	-	-	-
Hemoglobin increased	3	-	-	-	-	-
Lymphocyte count decreased	6	1	-	1	-	-
Neutrophil count decreased	14	5	-	1	-	-
Platelet count decreased	-	6	-	1	-	-
Weight gain	1	-	-	-	-	-
White blood cell decreased	6	1	-	-	-	-
Anorexia	3	1	-	1	-	-
Dehydration	2	-	-	3	-	-
Hyperglycemia	1	-	-	-	-	-
Hyperkalemia	1	-	-	-	-	-
Hyperuricemia	-	1	-	-	-	-
Hypoalbuminemia	2	-	-	-	-	-
Hyponatremia	3	-	-	-	-	-
Back pain	1	-	-	-	-	-
Generalized muscle weakness	4	-	-	-	-	-
Syncope	-	-	-	1	-	-
Confusion	1	-	-	-	-	-
Insomnia	1	-	-	-	-	-
Dyspnea	2	-	-	-	-	-
Hypoxia	1	-	-	-	-	-
Pleural effusion	1	-	-	-	-	-
Pulmonary edema	1	-	-	-	-	-
Acute kidney injury	1	-	-	-	-	-
Chronic kidney disease	1	-	-	-	-	-
Hematuria	1	-	-	-	-	-
Urinary retention	1	-	-	-	-	-
Urinary tract obstruction	1	-	-	-	-	-
Hypertension	4	1	-	8	-	-
Hypotension	-	-	-	-	1	-
Peripheral ischemia	-	-	-	1	-	-
Thromboembolic event	1	-	-	-	-	-
WORST DEGREE	25	10	1	16	1	-

Efficacy

In the sunitinib plus gemcitabine arm, 23 eligible patients were enrolled and treated in the first stage of the study. There were 6 responses, which met the threshold to move to the second stage, where an additional 22 eligible patients were enrolled and treated. In all, there was 1 complete response (CR) and 8 partial responses (PR) among the 45 eligible patients in the S+G arm for a response rate (80% two-stage confidence interval) of 20% [0.13, 0.31]. Interestingly, among the 9 patients with CR or PR, 6 were considered as having poor-risk disease, where the criteria for it included >50% sarcomatoid histology. With 1 CR and 8 PRs identified out of 45 eligible patients in the second stage (RR of 20%), the S+G arm did not meet the predetermined threshold for success of 30%.

The sunitinib alone arm had 1 CR, and 5 PRs, a response rate (80% two-stage confidence interval) of 16% [0.07, 0.25]. This RR met the predetermined threshold for success in this arm. In addition to the confirmed responses by RECIST, 6/37 patients had stable disease (SD). This element of response and stability as a part of disease control seemed to occur across all risk groups. However, these results are limited by the small sample size.

The waterfall plot describes all 69 patients with measurable disease (Figure 1). Of those, 37 had some degree of decrease in tumor size, 23 of whom were on S+G arm. Of the 23 from S+G arm, 1 had CR, 8 had PR, 8 had SD, 4 had PD, and 2 were unevaluable as the best response. 4 patients that had a PD either had new lesions on top of the reduction in tumor size or did not have confirmatory scans to confirm the decrease in the tumor size. Response rates by arm are detailed in Table 3. Responses were also examined by percent of sarcomatoid features in each arms. Patients with greater than 20% sarcomatoid features were considered to have a high proportion of aberrant histology. Thirty patients receiving S+G were noted to have high proportion of sarcomatoid features, but only 5/30 patients or 17% had a response in that group. Twenty-three patients in the sunitinib alone arm were noted to have high level of sarcomatoid features but had a low response (3/23 or 13%). Patients in both arms with PS of 2 had worse outcomes (data not shown).

Figure 1.

Waterfall plot of responses by treatment Arm in Patients on ECOG-ACRIN E1808.

Note: 4 bars with diagonal slashes are the patients with unevaluable responses and one patient with both vertical and horizontal lines is an ineligible patient (due to RT being completed ≤ 2 weeks prior to registration).

Table 3 –

Response rate by arm

	Arm A Sunitinib +gemcitabine N (%)	Arm B Sunitinib N (%)
Complete Response	1 (2.2)	1 (2.7)
Partial Response	8 (17.8)	5 (13.5)
Stable Disease	14 (31.1)	6 (16.2)
Progressive Disease	16 (35.6)	17 (45.9)
Unevaluable	6 (13.3)	8 (21.6)
Total	45	37

Survival analysis, utilizing Cox proportional hazards model, was used to explore both progression free survival (PFS) as well as overall survival (OS) of the two arms of the study. PFS and OS were both key secondary endpoints of the study. For arm A (S+G), the median PFS (90% CI) was 4.5 months (3.0, 5.8) and for arm B (S), median PFS 3.6 months (3.0, 7.7) (Figure 2). The median OS (90% CI) for arm A (S+G) was 9.4 months (7.4, 13.0) versus 7.8 months (7.0, 13.7) months for arm B (S) (Figure 3). Neither PFS nor OS showed significant differences between the two arms of the study when assessed by log-rank test.

Figure 2.

Progression free survival of patients by treatment Arm on ECOG-ACRIN E1808.

Figure 3.

Overall survival of patients by treatment Arm on ECOG-ACRIN E1808

Conclusions

Aberrant histology with sarcomatoid features or other non-clear cell types tend to have more aggressive and deadly courses of disease. This ECOG-ACRIN E1808 study was designed to assess the effectiveness and safety of the VEGF anti-angiogenic based therapy sunitinib plus or minus cytotoxic chemotherapy in sarcomatoid renal cell carcinoma. The study built upon earlier data suggesting cytotoxic therapy was of benefit in this particular histology ²³. While this study showed that there were no clinically significant differences between sunitinib alone and sunitinib plus gemcitabine combination therapy in sarcomatoid renal cell carcinoma, the sunitinib arm met its defined primary outcome, supporting the inclusion of VEGFi therapy in modern combination therapies. There was notable stability of disease in a number of patients in the sunitinib plus gemcitabine arm. Though not meeting its primary endpoint (30% RR, CI), the combination of sunitinib with gemcitabine had a higher response rate than sunitinib monotherapy. Given the responses seen with the addition of gemcitabine chemotherapy to sunitinib in the poor-risk patients, a chemotherapy approach with immune checkpoint inhibition, as is used in non-small cell lung cancer, could be explored as a direction in sarcomatoid RCC with these defined worst risk features.

Certainly, the overall therapeutic landscape for renal cell carcinoma has evolved dramatically over the last decade²⁵. Since design and completion of this E1808 study there have been numerous VEGFi and immunotherapy-based combinations approved for the 1^st line setting ^1-4. These have consistently reshaped guidelines for first-line and subsequent treatment of clear-cell RCC histology, with enhanced response rates and median overall survival data over time. Various combinations of these agents have also led to more effective therapies in patients with sarcomatoid histology ^26-28. Interestingly, as alluded to earlier, many of these histologies with sarcomatoid or rhabdoid features have responded well to checkpoint inhibitors ^29-31. The availability of these effective and potentially less harsh therapies has the potential to revolutionize care of the kidney cancer patient with aberrant histologies including those with sarcomatoid histology

In conclusion, this study is the largest and first randomized trial of VEGF inhibition combined with cytotoxic therapies for sarcomatoid RCC patients. The combination of sunitinib and gemcitabine did not meet the predetermined threshold for success. The majority of patients came off study due to progression or adverse events. Additional studies with novel therapeutics in sarcomatoid and other aggressive histologies may likely provide more potential for therapeutic benefit in these patients.

Key Points:

--Sarcomatoid RCC is an aggressive and rapidly growing histological variant with poor outcomes

--Prior data has shown limited effective options exist for this variant

--This demonstrates that cytotoxic chemotherapy in addition to VEGF inhibitors has limited effect and efficacy in these patients

--Novel therapeutic options will need to be examined in this subset of renal cell carcinoma patients

Clinical Practice Points.

Sarcomatoid renal cancer is an aggressive and rapidly growing histological variant with poor outcomes. A previous clinical trial of gemcitabine and doxorubicin demonstrated efficacy in some patients with this histology. We performed a cooperative group trial of sunitinib versus sunitinib and gemcitabine to ascertain if use of combination cytotoxic therapy with a vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGFR-TKI) improved benefit in patients with sarcomatoid renal cancer. We found that sunitinib alone had some benefit but that gemcitabine in addition to sunitinib had limited efficacy in these patients. After the trial was conducted, immune checkpoint inhibitor therapies were demonstrated to have some efficacy. Novel therapeutic options such as combining cytotoxic therapy or additional immune perturbation will need further study in this subset of renal cell carcinoma patients.