The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Sunitinib Alone or after Nephrectomy

in Metastatic Renal-Cell Carcinoma
A. Méjean, A. Ravaud, S. Thezenas, S. Colas, J.-B. Beauval, K. Bensalah,
L. Geoffrois, A. Thiery‑Vuillemin, L. Cormier, H. Lang, L. Guy, G. Gravis,
F. Rolland, C. Linassier, E. Lechevallier, C. Beisland, M. Aitchison, S. Oudard,
J.-J. Patard, C. Theodore, C. Chevreau, B. Laguerre, J. Hubert, M. Gross‑Goupil,
J.-C. Bernhard, L. Albiges, M.-O. Timsit, T. Lebret, and B. Escudier​​

A BS T R AC T

BACKGROUND
Cytoreductive nephrectomy has been the standard of care in metastatic renal-cell The authors’ full names, academic de-
carcinoma for 20 years, supported by randomized trials and large, retrospective grees, and affiliations are listed in the
Appendix. Address reprint requests to
studies. However, the efficacy of targeted therapies has challenged this standard. Dr. Escudier at the Department of Medi-
We assessed the role of nephrectomy in patients with metastatic renal-cell carci- cal Oncology, Gustave Roussy Institute,
noma who were receiving targeted therapies. 114 rue Edouard Vaillant, 94800 Villejuif,
France, or at ­escudier@​­gustaveroussy​.­fr.
METHODS Drs. Méjean and Ravaud contributed
In this phase 3 trial, we randomly assigned, in a 1:1 ratio, patients with confirmed equally to this article.
metastatic clear-cell renal-cell carcinoma at presentation who were suitable candi- This article was published on June 3, 2018,
dates for nephrectomy to undergo nephrectomy and then receive sunitinib (stan- at NEJM.org.
dard therapy) or to receive sunitinib alone. Randomization was stratified accord- DOI: 10.1056/NEJMoa1803675
ing to prognostic risk (intermediate or poor) in the Memorial Sloan Kettering Copyright © 2018 Massachusetts Medical Society.

Cancer Center prognostic model. Patients received sunitinib at a dose of 50 mg

daily in cycles of 28 days on and 14 days off every 6 weeks. The primary end point
was overall survival.
RESULTS
A total of 450 patients were enrolled from September 2009 to September 2017. At
this planned interim analysis, the median follow-up was 50.9 months, with 326
deaths observed. The results in the sunitinib-alone group were noninferior to
those in the nephrectomy–sunitinib group with regard to overall survival (strati-
fied hazard ratio for death, 0.89; 95% confidence interval, 0.71 to 1.10; upper
boundary of the 95% confidence interval for noninferiority, ≤1.20). The median
overall survival was 18.4 months in the sunitinib-alone group and 13.9 months in
the nephrectomy–sunitinib group. No significant differences in response rate or
progression-free survival were observed. Adverse events were as anticipated in each
group.
CONCLUSIONS
Sunitinib alone was not inferior to nephrectomy followed by sunitinib in patients
with metastatic renal-cell carcinoma who were classified as having intermediate-
risk or poor-risk disease. (Funded by Assistance Publique–Hôpitaux de Paris and
others; CARMENA ClinicalTrials.gov number, NCT00930033.)

n engl j med nejm.org 1
The New England Journal of Medicine
Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

K
idney cancer accounts for 5% of associated with improved survival versus initial
all cancers in men and 3% of all cancers targeted therapy.11
in women, and approximately 15% of The Surgery Time (SURTIME) prospective study
these are metastatic at diagnosis.1 More than was designed to investigate the role of initial
15 years ago, randomized, controlled trials showed nephrectomy as compared with deferred nephrec-
prolonged survival with initial nephrectomy, as tomy and showed longer overall survival with
compared with immunotherapy alone, among deferred nephrectomy than with initial nephrec-
patients with metastatic renal-cell cancer,2,3 and tomy; however, the study was underpowered for
nephrectomy became the standard of care for the evaluation of the survival end point because
patients who present with metastatic disease.4 of poor recruitment.12,13 In the CARMENA (Can-
Early retrospective analyses of prognostic factors cer du Rein Metastatique Nephrectomie et Anti-
that have been associated with survival in pa- angiogéniques) trial, we aimed to test the bene-
tients with metastatic renal-cell carcinoma simi- fit of initial nephrectomy followed by targeted
larly suggested an overall survival benefit with therapy in patients with metastatic kidney can-
cytoreductive nephrectomy.5,6 cer against the benefits provided by targeted
Since 2005, the treatment options have rapidly therapy alone.
expanded owing to the introduction of therapies
targeted to the molecular mechanisms underly- Me thods
ing renal-cancer carcinogenesis.7 Inhibitors of vas-
cular endothelial growth factor (VEGF) signal- Trial Oversight
ing, including the VEGF receptor tyrosine kinase The trial was designed under the auspices of the
inhibitors sunitinib and pazopanib and the anti- Cancerology Committee of the French Associa-
VEGF monoclonal antibody bevacizumab, are now tion of Urology and the Urogenital Tumors Study
standard first-line treatment options for favorable- Group and was sponsored by Assistance Pub-
risk and intermediate-risk metastatic renal-cell lique–Hôpitaux de Paris. The protocol (available
carcinoma, whereas the mammalian target of with the full text of this article at NEJM.org) was
rapamycin (mTOR) inhibitor temsirolimus is wide- approved by the French National Agency for the
ly recommended for patients who have renal-cell Safety of Medicines and Health Products (Agence
carcinoma and poor prognostic risk.8 Several Nationale de Sécurité du Médicament et des
other agents targeting the VEGF receptor, Produits de Santé).
mTOR, c-MET, or the interaction of the immune The sponsor and their representatives col-
checkpoint inhibitor programmed death 1 recep- lected and analyzed the data. All the authors had
tor with its ligand have shown efficacy and are full access to the trial data. Trial oversight was
current treatment options for patients with renal- provided by the first and second authors and by
cell carcinoma.8 an independent data and safety monitoring
The role of nephrectomy in treating meta- board (see the Supplementary Appendix, avail-
static renal-cell carcinoma in the era of targeted able at NEJM.org). Sunitinib was purchased at
therapy has been brought into question, since full cost from Pfizer per standard pharmacy
there is limited level 1 evidence directly compar- practice at each institute. Pfizer was not involved
ing the benefit of nephrectomy with targeted in the design or implementation of the protocol,
therapy.9 Multiple retrospective studies have in- the analysis of the data, or the preparation or
vestigated the relative benefits of nephrectomy review of the manuscript. Four of the authors
and targeted therapies, and these have supported wrote the first draft of the manuscript, and
a benefit of nephrectomy in patients who receive medical writing and editorial assistance was
targeted therapy. A systematic meta-analysis of funded by the sponsor. All the authors made the
cohort studies assessing the effect of nephrec- decision to submit the manuscript for publica-
tomy in patients receiving targeted therapy showed tion. The authors vouch for the accuracy and
that nephrectomy before targeted therapy was completeness of the data and the analyses and
associated with longer overall survival than tar- for the adherence of the trial to the protocol.
geted therapy alone,10 and a U.S. health database The conduct of the trial conformed with the
analysis showed that initial nephrectomy was International Conference on Harmonisation E6

2 n engl j med nejm.org

The New England Journal of Medicine

Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Sunitinib Alone or after Nephrectomy

guidelines for Good Clinical Practice and the one or two prognostic factors were classified as
principles of the Declaration of Helsinki. All the having intermediate-risk disease and those with
patients provided written informed consent before three or more were classified as having poor-risk
undergoing any trial procedures. disease (Table S1 in the Supplementary Appendix).
Nephrectomy was performed within 28 days
Patients after randomization, according to the normal
Eligible patients were adults (≥18 years of age) procedures of the institute. In the sunitinib-only
with clear-cell renal-cell carcinoma confirmed group, sunitinib treatment was initiated within
on mandatory biopsy and documented meta- 21 days after randomization and was given at an
static disease. Patients were required to have an initial dose of 50 mg daily in cycles of 28 days
Eastern Cooperative Oncology Group (ECOG) on followed by 14 days off every 6 weeks. In the
performance-status score of 0 or 1, an absence nephrectomy–sunitinib group, sunitinib treatment
of brain metastases or treated brain metastases was initiated between 3 and 6 weeks after ne-
without recurrence 3 weeks after treatment, and phrectomy. Dose reductions or interruptions of
acceptable organ function. ECOG scores range sunitinib treatment were permitted to manage
from 0 to 5, with higher numbers indicating adverse events. After recruitment and randomiza-
greater disability and a score of 5 indicating tion, each patient was followed for a minimum
death. Patients had to be suitable candidates for of 2 years.
nephrectomy and eligible for treatment with
sunitinib. Patients were assessed by the treating End Points and Assessments
urologist before enrollment regarding the feasi- The primary end point was overall survival,
bility of primary tumor resection; in cases in which was defined as the time from randomiza-
which there was doubt regarding feasibility, tion until death from any cause or until the date
computed tomographic (CT) scans of the abdo- of last contact for living patients. Secondary end
men were referred to the steering committee for points included investigator-assessed progression-
a decision. Patients were excluded if they had free survival, the objective response rate, clinical
received previous systemic treatment for kidney benefit (see below), adherence to treatment, ne-
cancer (including VEGF-targeted therapy) or anti- phrectomy in the sunitinib-only group, post­
coagulants or if they had any medical condition, operative morbidity and mortality, and safety.
including cardiovascular disease, that ruled them Progression-free survival was calculated, in
out as candidates for treatment. months, from the date of randomization to the
date of progression or the start date of a second-
Trial Design line treatment. Events in the analysis of progres-
In this prospective, multicenter, open-label, ran- sion-free survival included progression in patients
domized, phase 3 trial, patients were randomly undergoing treatment or during follow-up, the
assigned in a 1:1 ratio to undergo nephrectomy start of a new line of treatment, and death from
followed by sunitinib treatment or to receive cancer-related causes. Tumor response was as-
sunitinib alone. Randomization was stratified sessed according to the Response Evaluation
according to risk group (classified according to Criteria in Solid Tumors, version 1.1, and consid-
the Memorial Sloan Kettering Cancer Center ered all lesions (target, nontarget, and primary).
[MSKCC] prognostic model) and center.5 The The objective response rate was defined as the
MSKCC prognostic factors were: a Karnofsky percentage of patients with a complete response
performance status score of less than 80 (on a or partial response. Clinical benefit was defined
scale from 0 to 100, with lower scores indicating as the percentage of patients with a complete
greater disability), a lactate dehydrogenase level response, partial response, or stable disease for
of 1.5 times the upper limit of the normal range, more than 12 weeks. Postoperative morbidity
a hemoglobin level that was less than the lower was evaluated according to the Clavien–Dindo
limit of the normal range, a corrected serum classification of surgical complications14 (Table S2
calcium level of more than 10 mg per deciliter in the Supplementary Appendix), and postopera-
(2.5 mmol per liter), and a time from diagnosis tive mortality was evaluated as the percentage of
to treatment of less than 1 year. Patients with deaths in the 30 days after nephrectomy. Adverse

n engl j med nejm.org 3
The New England Journal of Medicine
Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

events were evaluated according to the Common sunitinib-alone group who had received sunitinib;
Terminology Criteria for Adverse Events, version and the second included only patients in the
3.0, of the National Cancer Institute. Radio- nephrectomy–sunitinib group who had under-
graphic tumor evaluation (CT scan or magnetic gone nephrectomy and also received sunitinib
resonance imaging [MRI] of the thorax, abdo- and patients in the sunitinib-alone group who
men, and pelvis) was performed after every two had received sunitinib.
cycles of sunitinib treatment; CT or MRI scans
of the head were performed only if clinically R e sult s
indicated.
Patients
Statistical Analysis Between September 23, 2009, and September 8,
To assess the hypothesis that nephrectomy is not 2017, a total of 450 patients were enrolled from
necessary in patients presenting with metastatic 79 centers in France and from other centers in
kidney cancer, we considered that treatment with Europe (425 patients from France, 14 from the
sunitinib alone from the outset would be consid- United Kingdom, 10 from Norway, and 1 from
ered to be clinically acceptable if the upper Sweden) (Table S3 in the Supplementary Appen-
boundary of the 95% confidence interval for dix). Patients were randomly assigned in a 1:1
the hazard ratio for death was 1.20 or less (non- ratio to undergo nephrectomy and then receive
inferiority margin). The trial was designed to sunitinib treatment (226 patients) or to start
have 80% power at a one-sided significance level treatment with sunitinib alone immediately (224
of 5%. To show noninferiority, we planned to patients) (Fig. 1). In the nephrectomy–sunitinib
enroll 576 patients in order to observe 456 deaths, group, 16 patients (7.1%) did not undergo ne-
on the basis of an expected recruitment of 12 phrectomy and 40 (17.7%) never received suni-
patients per month over a period of 48 months. tinib; in the sunitinib-alone group, 11 patients
This allowed for two interim analyses — one (4.9%) never received sunitinib and 38 (17.0%)
after the observation of 152 deaths and another underwent subsequent nephrectomy a median of
after 304 deaths — and for a final analysis to be 11.1 months after randomization for the control
scheduled either 80 months after the initiation of symptoms. Details regarding the nephrecto-
of the trial or 32 months after the enrollment mies are provided in Table S6 in the Supplemen-
of the last patient. O’Brien–Fleming sequential tary Appendix.
boundaries were used for decisions regarding At the time of data cutoff (December 12,
early halting of the trial. The second interim 2017), the median follow-up of the patients was
analysis after the observation of 326 events at 50.9 months overall (95% confidence interval
the cutoff date of December 12, 2017, reported [CI], 44.0 to 56.9; range, 0.0 to 86.6). The major-
here, showed that the O’Brien–Fleming bound- ity of patients were men (74.7%), the median age
ary was not reached, so the trial was allowed to of patients in the intention-to-treat population
continue recruitment. However, in parallel to this was 62 years (range, 30 to 87), and 56.0% of the
second interim analysis, the sponsor made the patients overall had an ECOG performance-
decision to close the trial early because of slow status score of 0. The demographic and clinical
recruitment, and the steering committee consid- characteristics of the patients at baseline were
ered the results from this interim analysis suf- well balanced between the treatment groups
ficient to meet the objectives of the trial. (Table 1). In the nephrectomy–sunitinib group,
The rates and 95% confidence intervals for 55.6% of the patients were in the MSKCC inter-
the analyses of overall survival and progression- mediate-risk group and 44.4% were in the poor-
free survival were estimated by the Kaplan–Meier risk group; in the sunitinib-alone group, the
method in the intention-to-treat population. Tu- corresponding values were 58.5% and 41.5%.
mor response and safety data were analyzed in Sunitinib treatment was stopped in 67.1% of
patients who received sunitinib. Two per-protocol the patients because of disease progression, in
analyses were performed: the first included only 13.0% because of toxic effects, and in 6.2% be-
patients in the nephrectomy–sunitinib group who cause of death. Subsequent anticancer therapies
had undergone nephrectomy and patients in the were received by similar percentages of patients

4 n engl j med nejm.org

The New England Journal of Medicine

Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Sunitinib Alone or after Nephrectomy

450 Patients were enrolled and underwent

randomization (intention-to-treat population)

226 Were assigned to undergo cytoreductive 224 Were assigned to receive sunitinib alone
nephrectomy and receive sunitinib 8 Had deviation from inclusion criteria
6 Had deviation from inclusion criteria

11 Did not receive sunitinib

40 Did not receive sunitinib

186 Were included in the safety population 213 Were included in the safety population

38 Underwent subsequent
16 Did not undergo surgery nephrectomy, including
3 Withdrew consent 3 who were not treated
2 Were lost to follow-up with sunitinib
2 Were lost to follow-up
114 Received further lines
of therapy 115 Received further lines
of therapy

165 Died 161 Died

Figure 1. Randomization, Treatment, and Follow-up of the Patients.

The numbers of patients who died were based on the intention-to-treat population. The patients who did not undergo
surgery (in the nephrectomy–sunitinib group) or who underwent subsequent nephrectomy (in the sunitinib-alone
group), who withdrew consent, who were lost to follow-up, or who received further lines of therapy were all included
in the analyses of overall survival and progression-free survival.

in the nephrectomy–sunitinib group and the ratio did not exceed the fixed noninferiority limit
sunitinib-alone group (50.4% and 51.3%, respec- (1.20). Thus, sunitinib alone was not inferior to
tively), the most common being everolimus (in nephrectomy followed by sunitinib. In both the
21.9% and 31.3%) and axitinib (in 23.7% and intermediate-risk and poor-risk groups of pa-
25.2%) (Table S7 in the Supplementary Appendix). tients, the median overall survival was longer in
the sunitinib-alone group than in the nephrec-
Efficacy tomy–sunitinib group (23.4 vs. 19.0 months in
At the time of data cutoff, 326 deaths had oc- the intermediate-risk subgroup and 13.3 vs. 10.2
curred, of which 91.0% were cancer-related. In the months in the poor-risk group). In the interme-
intention-to-treat population, patients in the suni- diate-risk population, the hazard ratio for death
tinib-alone group had a longer median overall in the sunitinib-alone group, as compared with
survival than those in the nephrectomy–suni- the nephrectomy–sunitinib group, was 0.92 (95%
tinib group (18.4 months [95% CI, 14.7 to 23.0] CI, 0.68 to 1.24), and in the poor-risk population,
vs. 13.9 months [95% CI, 11.8 to 18.3]) (Fig. 2A). the hazard ratio was 0.86 (95% CI, 0.62 to 1.17).
The hazard ratio for death in the analysis of At the time of data cutoff, 390 events had
overall survival, stratified according to MSKCC been observed for the analysis of progression-
risk score, was 0.89 (95% CI, 0.71 to 1.10). At free survival; 86.7% of the events were disease
this planned interim analysis, the upper bound- progression, death from disease progression, or
ary of the 95% confidence interval for the hazard the start of a new line of treatment. Although

n engl j med nejm.org 5
The New England Journal of Medicine
Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 1. Demographic and Clinical Characteristics of the Patients at Baseline.*

Nephrectomy–Sunitinib Sunitinib Alone

Characteristic (N = 226) (N = 224)
Median age (range) — yr 63 (33–84) 62 (30–87)
Male sex — no. (%) 169 (74.8) 167 (74.6)
MSKCC risk category — no./total no. (%)†
Intermediate risk 125/225 (55.6) 131/224 (58.5)
Poor risk 100/225 (44.4) 93/224 (41.5)
ECOG performance-status score — no. (%)‡
0 130 (57.5) 122 (54.5)
1 96 (42.5) 102 (45.5)
Fuhrman grade of renal-cell carcinoma — no./total no. (%)§
1 or 2 77/150 (51.3) 82/156 (52.6)
3 or 4 73/150 (48.7) 74/156 (47.4)
Tumor–node–metastasis stage — no./total no. (%)¶ 71/207 (34.3) 50/194 (25.8)
Tumor stage
T1 5/67 (7.5) 7/49 (14.3)
T2 13/67 (19.4) 13/49 (26.5)
T3 or 4 47/67 (70.1) 25/49 (51.0)
Tx 2/67 (3.0) 4/49 (8.2)
Node stage
N0 23/66 (34.8) 18/49 (36.7)
N1 13/66 (19.7) 6/49 (12.2)
N2 7/66 (10.6) 13/49 (26.5)
Nx 23/66 (34.8) 12/49 (24.5)
Median primary tumor size (range) — mm 88 (6–200) 86 (12–190)
Median no. of metastatic sites (range) 2 (1–5) 2 (1–5)
Median tumor burden (range) — mm‖ 140 (23–399) 144 (39–313)
Location of metastases — no./total no. (%)
Lung 172/217 (79.3) 161/221 (72.9)
Bone 78/217 (35.9) 82/221 (37.1)
Lymph nodes 76/217 (35.0) 86/221 (38.9)
Other 78/217 (35.9) 90/221 (40.7)

* The characteristics of the patients at baseline were well balanced between the treatment groups.
† The Memorial Sloan Kettering Cancer Center (MSKCC) prognostic factors regarding risk were: a Karnofsky performance-
status score of less than 80 (on a scale from 0 to 100, with lower scores indicating greater disability), a lactate dehydro-
genase level of 1.5 times the upper limit of the normal range, a hemoglobin level that was less than the lower limit of
the normal range, a corrected serum calcium level of more than 10 mg per deciliter (2.5 mmol per liter), and a time
from diagnosis to treatment of less than 1 year. Patients with one or two prognostic factors were classified as having
intermediate-risk disease and those with three or more were classified as having poor-risk disease.
‡ Eastern Cooperative Oncology Group (ECOG) performance-status scores range from 0 to 5, with higher scores indicat-
ing greater disability and a score of 5 indicating death.
§ The Fuhrman grade of renal-cell carcinoma is assessed on a scale of 1 to 4, with grade 1 indicating the least atypia and
grade 4 the most.
¶ The tumor–node–metastasis stage was determined according to the criteria of the Union for International Cancer
Control TNM Classification of Malignant Tumours.15 Metastasis was an eligibility criterion in all patients.
‖ Tumor burden was assessed according to the Response Evaluation Criteria in Solid Tumors, version 1.1.

6 n engl j med nejm.org

The New England Journal of Medicine

Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Sunitinib Alone or after Nephrectomy

radiographic imaging of tumors was not man-

Nephrectomy–sunitinib Sunitinib alone
datory before the initiation of sunitinib in the
nephrectomy–sunitinib group, 223 patients had A Overall Survival
radiographic imaging data of target lesions at 100
baseline, and of the 176 patients who received 90

Patients Who Were Alive (%)

sunitinib, 111 had radiographic imaging of the 80
70
target lesion performed at treatment initiation. 64.4
60
In the sunitinib-alone group, 222 of 223 patients 55.2
50
with data (99.6%) had a CT or MRI scan at base- 42.6
40
line, before the initiation of sunitinib alone. The 35.0 29.1
30
median progression-free survival was longer 25.9
20
among patients in the sunitinib-alone group 10
than among those in the nephrectomy–sunitinib 0
group (8.3 months [95% CI, 6.2 to 9.9] vs. 7.2 0 12 24 36 48 60 72 84 96
months [95% CI, 6.7 to 8.5]) (Fig. 2B). The haz- Months
ard ratio for progression or death, stratified ac- No. at Risk
cording to risk group, was 0.82 (95% CI, 0.67 to Nephrectomy– 226 110 61 40 19 11 4 1 0
sunitinib
1.00). The results of the per-protocol analyses Sunitinib alone 224 128 76 44 26 8 3 1 0
were consistent with those of the intention-to-
treat analysis (Tables S4 and S5 in the Supple- B Progression-free Survival
mentary Appendix). 100

The objective response rate was similar in the 90

Patients without an Event (%)

80
two trial groups (29.1% in the sunitinib-alone
70
group and 27.4% in the nephrectomy–sunitinib
60
group) (Table 2), although the rate of disease
50
control was nonsignificantly higher in the suni- 37.3
40
tinib-alone group than in the nephrectomy– 30 30.4
sunitinib group (74.6% vs. 61.8%) (Table 2). 20 14.5
3.7
Clinical benefit was observed in 47.9% of the 10 5.2
patients in the sunitinib-alone group, as com- 0
6.1
pared with 36.6% of those in the nephrectomy– 0 12 24 36 48 60 72
sunitinib group (P = 0.02) (Table 2). Months
Postoperative death in the month after ne- No. at Risk
Nephrectomy– 226 59 10 6 2 1 0
phrectomy was recorded in 4 patients in the sunitinib
nephrectomy–sunitinib group. Among patients Sunitinib alone 224 74 28 9 6 2 0
with postoperative complications in the nephrec-
tomy–sunitinib group (82 of 210 patients), com- Figure 2. Kaplan–Meier Estimates of Survival.
plications of Clavien–Dindo grade III or higher The landmark analyses of overall survival (Panel A) and progression-free
survival (Panel B) at 1, 2, and 3 years are indicated by the vertical dashed
occurred in 15.9% (13 of 82 patients).
lines, with rates in the two groups shown at those time points. The hori-
zontal dashed line in each panel indicates the median. Tick marks indicate
Exposure and Safety censored data.
The median duration of sunitinib treatment was
6.7 months (range, 1.4 to 67.2) in the nephrec-
tomy–sunitinib group and 8.5 months (range, Only severe (grade 3 or 4) adverse events are
0.9 to 63.7) in the sunitinib-alone group (P = 0.04). reported in this article because the safety profile
Dose reduction occurred in 57 of 186 patients of sunitinib is well established. Overall, 38.1% of
(30.6%) in the nephrectomy–sunitinib group and the patients had an adverse event of grade 3 or 4;
in 65 of 213 (30.5%) in the sunitinib-alone group. a total of 61 patients (32.8%) in the nephrectomy–
The majority of dose reductions were done to sunitinib group and 91 (42.7%) in the sunitinib-
manage adverse events. alone group reported adverse events of grade 3

n engl j med nejm.org 7
The New England Journal of Medicine
Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 2. Tumor Response Outcomes.*

Nephrectomy–Sunitinib Sunitinib Alone

Response (N = 186) (N = 213)
Best overall response — no./total no. (%)
Complete response 1/178 (0.6) 0/208
Partial response 50/178 (28.1) 62/208 (29.8)
Stable disease 64/178 (36.0) 97/208 (46.6)
Progression of disease 49/178 (27.5) 40/208 (19.2)
Could not be evaluated 14/178 (7.9) 9/208 (4.3)
Objective response rate — % (95% CI)† 27.4 (21.1–34.4) 29.1 (23.1–35.7)
Disease control rate — % (95% CI)‡ 61.8 (54.4–68.8) 74.6 (68.2–80.3)
Clinical benefit — no. (%)§ 68 (36.6) 102 (47.9)

* Tumor response was analyzed in patients who received sunitinib. Some patients could not be evaluated for tumor response
because of adverse events during treatment or deterioration of condition.
† Objective response was defined as a complete or partial response.
‡ Disease control was defined as a complete or partial response or stable disease.
§ Clinical benefit was defined as disease control beyond 12 weeks (P = 0.02 for this comparison).

or 4 (P = 0.04) (Table 3). The most common ad- These findings contrast with those of previous
verse events of grade 3 or 4 that were observed retrospective and database studies, which suggest­
in the 152 patients in the safety population (i.e., ed an overall survival benefit with nephrectomy
patients treated with sunitinib) who reported an in patients treated with targeted therapies.10,11,16
adverse event of grade 3 or 4 were asthenia (in Retrospective analyses such as these are subject
37 patients), the hand–foot syndrome (in 20), to selection bias; patients who are selected for
anemia (in 16), and neutropenia (in 15). Grade 3 nephrectomy are more likely to have more favor-
or 4 renal or urinary tract disorders occurred in able clinical characteristics, such as good perfor-
1 patient in the nephrectomy–sunitinib group and mance status and limited metastatic volume,
in 9 in the sunitinib-alone group (P = 0.051). For than those who are not selected for surgery, a
all other toxic effects, no significant difference factor that potentially contributes to survival dif-
was evident between the trial groups. ferences. Data from a large, retrospective study
by the International Metastatic Renal Cell Carci-
noma Database Consortium (IMDC) showed that
Discussion
patients undergoing nephrectomy had better prog-
Nephrectomy is currently recommended in pa- nostic profiles than patients whose diseased
tients with a good performance status and large kidney was not removed and that prognostic
primary tumors with limited volumes of meta- factors modified the effect of the operation on
static disease and is not recommended in patients survival.17 Patients with multiple risk factors did
with poor performance status,4 but there has not have a survival benefit with nephrectomy.10
been limited level 1 evidence to support the use In a National Cancer Database analysis involving
of nephrectomy in the context of targeted ther­ 4223 patients with metastatic renal-cell carci-
apy. Given the many approved options for sys- noma who were treated with nephrectomy and
temic targeted therapy that are now available, targeted therapy, overall survival rates were sig-
the reassessment of the role of surgery in dis- nificantly lower among patients who underwent
ease management is important. This trial showed nephrectomy before targeted therapy than among
that sunitinib alone was not inferior to nephrec- those who underwent nephrectomy after target-
tomy followed by sunitinib in patients with ed therapy.16 However, analyses of hospital reg-
metastatic renal-cell carcinoma who had been istries such as these are subject to selection bias
classified as having MSKCC intermediate-risk or regarding how clinicians opted to administer
poor-risk disease. targeted therapy (before or after nephrectomy).

8 n engl j med nejm.org

The New England Journal of Medicine

Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Sunitinib Alone or after Nephrectomy

Table 3. Summary of Severe Adverse Events in Sunitinib-Treated Patients.*

Nephrectomy–Sunitinib Sunitinib Alone

Event (N = 186) (N = 213)

no. of patients (%)

Any adverse event of grade 3 or 4† 61 (32.8) 91 (42.7)
Asthenia 16 (8.6) 21 (9.9)
Inflammation of mucosa 1 (0.5) 6 (2.8)
Edema 0 4 (1.9)
Neutropenia 5 (2.7) 10 (4.7)
Thrombocytopenia 7 (3.8) 5 (2.3)
Anemia 5 (2.7) 11 (5.2)
Hand–foot syndrome 8 (4.3) 12 (5.6)
Intratumoral hemorrhage 0 1 (0.5)
Pulmonary embolism 2 (1.1) 2 (0.9)
Severe high blood pressure 6 (3.2) 7 (3.3)
Left ventricular failure 0 1 (0.5)
Heart failure 0 1 (0.5)
Hepatitis 1 (0.5) 1 (0.5)
Liver failure 0 2 (0.9)
Severe hypothyroidism 3 (1.6) 1 (0.5)
Musculoskeletal or systemic disorder 2 (1.1) 5 (2.3)
Respiratory, thoracic, or mediastinal disorder 3 (1.6) 4 (1.9)
Renal or urinary tract disorder‡ 1 (0.5) 9 (4.2)
Gastrointestinal perforation 1 (0.5) 2 (0.9)
Seizure or convulsions 0 1 (0.5)
Other 34 (18.3) 47 (22.1)

* Shown are adverse events of grade 3 or 4 that were observed among patients who received sunitinib.
† P = 0.04.
‡ P = 0.051.

Avoiding surgery can provide other benefits applicable to patients with a poor performance
for patients. Initial nephrectomy can delay the status, minimal primary tumor burden, and high
start of systemic targeted therapies that have volumes of metastatic disease, because these
shown a survival benefit, and patients may die patients are not generally recommended to under­
before receiving such therapies.18 Avoiding ne- go nephrectomy.4 The use of MSKCC risk groups,
phrectomy also avoids surgical complications which were the risk groups in common use at
involving blood transfusions, further operations, the time the trial was launched, is an unavoid-
or intensive care, which may also delay systemic able limitation of this analysis, since these are
therapy.19 In addition, there is uncertainty about not as relevant as IMDC risk groups in the era of
which patients are appropriate candidates for targeted therapy. In addition, the inclusion of
nephrectomy.8 patients with minimal tumor burden could have
A potential limitation of this trial is that en- resulted in different survival outcomes. Because
rolled patients were appropriate candidates for this was a noninferiority trial, the results may
nephrectomy in the opinion of the treating underestimate the benefit of nephrectomy. An-
urologist; therefore, the results are not generally other limitation of this trial is the recruitment of

n engl j med nejm.org 9
The New England Journal of Medicine
Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

fewer patients than planned (450 patients rather tients with disease features indicating poor risk,
than 576), which reduced the statistical power. and the findings confirm current treatment
However, the trend in longer overall survival and guideline recommendations for systemic target-
progression-free survival among patients who ed therapy in patients with poor-risk metastatic
did not undergo nephrectomy suggests that our renal-cell carcinoma.
conclusion is correct. Finally, the exclusion, at Although nephrectomy may have a role in con-
the investigator’s discretion, of patients with low trolling symptoms in some patients with meta-
metastatic burden could be considered to result static renal-cell carcinoma, as suggested by retro-
in a potential bias, and this situation may have spective studies,11,17 there is no “one size fits all”
contributed to the high proportion of patients approach. The multimodal approach of individ-
with features indicating poor prognostic risk and ualized treatment provides appropriate manage-
the relatively short overall survival that was ob- ment of metastatic renal-cell carcinoma. Data are
served in the trial. These patients are usually con- lacking regarding the role of nephrectomy be-
sidered to be good candidates for nephrectomy fore immune-checkpoint inhibitors in patients
followed by surveillance. with renal-cell carcinoma.
Sunitinib was the therapy chosen in the de- In conclusion, in this trial, sunitinib alone
sign of CARMENA on the basis of supporting was not inferior to nephrectomy followed by
evidence that was available at the time. Although sunitinib in patients with metastatic renal-cell
sunitinib and pazopanib are currently the most carcinoma who were in the MSKCC intermediate-
commonly used treatments in patients with risk or poor-risk groups.
metastatic renal-cell carcinoma with a good or Supported by Assistance Publique–Hôpitaux de Paris (Clinical
intermediate prognosis,4 two recent randomized Research and Innovation Delegation), by a grant (PHRC-K 2007)
from the Programme Hospitalier de Recherche Clinique Cancer–
trials involving patients with an intermediate Ministère de la Santé, and by an unrestricted grant from Pfizer.
or poor prognosis showed the superiority of the Disclosure forms provided by the authors are available with
c-MET inhibitor cabozantinib20 and immune the full text of this article at NEJM.org.
We thank the staff of the Unité de Recherche Clinique–Centre
checkpoint inhibitor combination therapy (nivolu­ d’Investigation Clinique Paris Descartes Necker–Cochin for the im-
mab plus ipilimumab)21 over sunitinib. These plementation, monitoring, and data management of the trial; the
agents will probably be incorporated as initial members of the data and safety monitoring board (see the Supple-
mentary Appendix); Lucy Smithers, Ph.D., of ApotheCom, for medi-
treatment options for patients with intermediate- cal and editing assistance with an earlier version of the manuscript;
risk or poor-risk disease. Our trial included pa- and all the investigators and patients who participated in the trial.

Appendix
The authors’ full names and academic degrees are as follows: Arnaud Méjean, M.D., Ph.D., Alain Ravaud, M.D., Ph.D., Simon Theze-
nas, Ph.D., Sandra Colas, M.D., Jean‑Baptiste Beauval, M.D., Karim Bensalah, M.D., Ph.D., Lionnel Geoffrois, M.D., Antoine
Thiery‑Vuillemin, M.D., Ph.D., Luc Cormier, M.D., Ph.D., Hervé Lang, M.D., Ph.D., Laurent Guy, M.D., Ph.D., Gwenaelle Gravis, M.D.,
Frederic Rolland, M.D., Claude Linassier, M.D., Ph.D., Eric Lechevallier, M.D., Ph.D., Christian Beisland, M.D., Ph.D., Michael Aitchi-
son, M.D., Stephane Oudard, M.D., Ph.D., Jean‑Jacques Patard, M.D., Ph.D., Christine Theodore, M.D., Christine Chevreau, M.D.,
Brigitte Laguerre, M.D., Jacques Hubert, M.D., Marine Gross‑Goupil, M.D., Ph.D., Jean‑Christophe Bernhard, M.D., Ph.D., Laurence
Albiges, M.D., Ph.D., Marc‑Olivier Timsit, M.D., Ph.D., Thierry Lebret, M.D., Ph.D., and Bernard Escudier, M.D.
The authors’ affiliations are as follows: the Departments of Urology (A.M., M.-O.T.) and Medical Oncology (S.O.), Hôpital Européen
Georges-Pompidou, Assistance Publique–Hôpitaux de Paris (AP-HP), Université Paris–Descartes, and Paris Descartes Necker–Cochin
Clinical Research Unit, AP-HP (S.C.), Paris, the Department of Medical Oncology, Bordeaux University Hospital (A.R.), the Department
of Medical Oncology, Hôpital Saint André, Centre Hospitalier Universitaire (CHU) de Bordeaux (M.G.-G.), and the Department of Urol-
ogy, CHU de Bordeaux and Université de Bordeaux (J.-C.B.), Bordeaux, Biometrics Unit, Cancer Institute of Montpellier, University of
Montpellier, Montpellier (S.T.), the Department of Urology, CHU Rangueil (J.-B.B.), and the Department of Medical Oncology, Institut
Universitaire du Cancer Toulouse–Oncopole (C.C.), Toulouse, the Department of Urology, University of Rennes (K.B.), and the Depart-
ment of Medical Oncology, Centre Eugene Marquis (B.L.), Rennes, the Department of Medical Oncology, Institut de Cancerologie de
Lorraine, Vandoeuvre lès Nancy (L. Geoffrois), the Department of Medical Oncology, CHU Besançon, Oncologie, and Université de
Franche-Comte, INSERM Unité Mixte de Recherche (UMR) 1098, Structure Fédérative de Recherche Ingénierie et Biologie Cellulaire
et Tissulaire, Besançon (A.T.-V.), the Department of Urology, CHU François Mitterrand, Dijon (L.C.), the Department of Urology, CHU
Strasbourg, Translational Medicine Federation Strasbourg, Strasbourg (H.L.), the Department of Urology, Gabriel Montpied Hospital,
and Clermont Auvergne University, Clermont-Ferrand (L. Guy), the Department of Medical Oncology, Centre de Recherche en Cancer-
ologie de Marseille, INSERM UMR 1068, Centre National de la Recherche Scientifique UMR 7258 and Institut Paoli-Calmettes (G.G.),
and the Department of Urology, Hôpital la Conception (E.L.), Aix Marseille Université, Marseille, the Department of Medical Oncology,
Institut de Cancerologie de l’Ouest, Nantes (F.R.), the Department of Medical Oncology, CHU Bretonneau, and the Department of
Medicine, Université François Rabelais, Tours (C.L.), the Department of Urology, Mont de Marsan General Hospital, Mont de Marsan
(J.-J.P.), the Departments of Medical Oncology (C.T.) and Urology (T.L.), Hôpital Foch, Suresnes, the Department of Urology, Imagerie
Adaptative Diagnostique et Interventionnelle INSERM Unité 1254, CHU de Nancy, Brabois (J.H.), the Department of Medical Oncology,

10 n engl j med nejm.org

The New England Journal of Medicine

Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Sunitinib Alone or after Nephrectomy

Institut Gustave Roussy and Université Paris-Saclay, Villejuif (L.A., B.E.), and Université Versailles, St.-Quentin-en-Yvelines (T.L.) — all
in France; the Department of Urology, Haukeland University Hospital (C.B.), and the Department of Clinical Medicine, University of
Bergen (C.B.), Bergen, Norway; and the Department of Urology, Royal Free Hospital, London (M.A.).

References
1. Siegel RL, Miller KD, Jemal A. Cancer a systematic review and critical evaluation. lignant tumours. 7th ed. Oxford, United
statistics, 2018. CA Cancer J Clin 2018;​68:​ Oncologist 2017;​22:​667-79. Kingdom:​Wiley-Blackwell, 2009.
7-30. 9. Culp SH. Cytoreductive nephrectomy 16. Hanna N, Sun M, Meyer CP, et al. Sur-
2. Flanigan RC, Salmon SE, Blumenstein and its role in the present-day period of vival analyses of patients with metastatic
BA, et al. Nephrectomy followed by inter- targeted therapy. Ther Adv Urol 2015;​7:​ renal cancer treated with targeted therapy
feron alfa-2b compared with interferon 275-85. with or without cytoreductive nephrec-
alfa-2b alone for metastatic renal-cell 10. García-Perdomo HA, Zapata-Copete tomy: a National Cancer Data Base study.
cancer. N Engl J Med 2001;​345:​1655-9. JA, Castillo-Cobaleda DF. Role of cyto­ J Clin Oncol 2016;​34:​3267-75.
3. Mickisch GH, Garin A, van Poppel H, reductive nephrectomy in the targeted 17. Heng DY, Wells JC, Rini BI, et al. Cyto-
de Prijck L, Sylvester R. Radical nephrec- therapy era: a systematic review and meta- reductive nephrectomy in patients with
tomy plus interferon-alfa-based immu- analysis. Investig Clin Urol 2018;​59:​2-9. synchronous metastases from renal cell
notherapy compared with interferon alfa 11. Bhindi B, Habermann EB, Mason RJ, carcinoma: results from the International
alone in metastatic renal-cell carcinoma: et al. Comparative survival following initial Metastatic Renal Cell Carcinoma Database
a randomised trial. Lancet 2001;​358:​966- cytoreductive nephrectomy versus initial Consortium. Eur Urol 2014;​66:​704-10.
70. targeted therapy for metastatic renal cell 18. Bhindi B, Habermann E, Mason R, et al.
4. Escudier B, Porta C, Schmidinger M, carcinoma. J Urol 2018 March 21 (Epub Survival following upfront cytoreductive
et al. Renal cell carcinoma: ESMO clinical ahead of print). nephrectomy versus targeted therapy for
practice guidelines for diagnosis, treat- 12. Stewart GD, Aitchison M, Bex A, et al. metastatic renal cell carcinoma. J Clin
ment and follow-up. Ann Oncol 2016;​27:​ Cytoreductive nephrectomy in the tyro- Oncol 2018;​36:​Suppl:​672. abstract.
Suppl 5:​v58-v68. sine kinase inhibitor era: a question that 19. Silberstein JL, Adamy A, Maschino AC,
5. Motzer RJ, Mazumdar M, Bacik J, may never be answered. Eur Urol 2017;​71:​ et al. Systematic classification and predic-
Berg W, Amsterdam A, Ferrara J. Survival 845-7. tion of complications after nephrectomy
and prognostic stratification of 670 pa- 13. Bex A, Mulders P, Jewett MAS, et al. in patients with metastatic renal cell car-
tients with advanced renal cell carcinoma. Immediate versus deferred cytoreductive cinoma (RCC). BJU Int 2012;​110:​1276-82.
J Clin Oncol 1999;​17:​2530-40. nephrectomy (CN) in patients with syn- 20. Choueiri TK, Halabi S, Sanford BL, et
6. Neves RJ, Zincke H, Taylor WF. Meta- chronous metastatic renal cell carcinoma al. Cabozantinib versus sunitinib as ini-
static renal cell cancer and radical nephrec- (mRCC) receiving sunitinib (EORTC 30073 tial targeted therapy for patients with meta-
tomy: identification of prognostic factors SURTIME). Eur Urol Suppl 2018;​ 17:​
(2):​ static renal cell carcinoma of poor or
and patient survival. J Urol 1988;​ 139:​ e1-e2. abstract. intermediate risk: the Alliance A031203
1173-6. 14. Clavien PA, Sanabria JR, Strasberg SM. CABOSUN Trial. J Clin Oncol 2017;​35:​
7. Choueiri TK, Motzer RJ. Systemic Proposed classification of complications of 591-7.
therapy for metastatic renal-cell carcino- surgery with examples of utility in chole- 21. Motzer RJ, Tannir NM, McDermott
ma. N Engl J Med 2017;​376:​354-66. cystectomy. Surgery 1992;​111:​518-26. DF, et al. Nivolumab plus ipilimumab ver-
8. Bamias A, Escudier B, Sternberg CN, 15. Sobin LH, Gospodarowicz MK, Witte­ sus sunitinib in advanced renal-cell carci-
et al. Current clinical practice guidelines kind C, eds. International Union against noma. N Engl J Med 2018;​378:​1277-90.
for the treatment of renal cell carcinoma: Cancer (UICC) TNM classification of ma- Copyright © 2018 Massachusetts Medical Society.

n engl j med nejm.org 11
The New England Journal of Medicine
Downloaded from nejm.org on June 3, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.