Clinical Study

Oncology Received: February 22, 2017

Accepted after revision: April 11, 2017
DOI: 10.1159/000475715
Published online: June 10, 2017

Outcomes in Melanoma Patients Treated

with BRAF/MEK-Directed Therapy or Immune
Checkpoint Inhibition Stratified by Clinical Trial
versus Standard of Care
Chloe Goldman a Jeremy Tchack a Eric M. Robinson a Sung Won Han b
       

Una Moran a David Polsky a, c Russell S. Berman d Richard L. Shapiro d

       

Patrick A. Ott e Iman Osman a Hua Zhong b Anna C. Pavlick b

       

Melissa Ann Wilson e   

a
The Ronald O. Perelman Department of Dermatology, b Department of Population Health, Environmental
   

Medicine, c Department of Pathology, d Division of Surgical Oncology, Department of Surgery, and

   

e
Department of Medicine, Division of Hematology and Oncology, Perlmutter Cancer Center, New York University
 

School of Medicine, New York, NY, USA

Keywords ferences between trial and commercial cohorts were as-

Metastatic melanoma · Immunotherapy · Targeted therapy · sessed. Results: Patients receiving commercial therapy (vs.
Immune-related adverse events · Adverse events on trial) had poorer prognostic features (i.e., brain metasta-
ses) and lower median overall survival (mOS) when assessed
across all treatments (9.2 vs. 17.5 months, p = 0.0027). While
Abstract toxicity within trial and commercial cohorts did not differ,
Objectives: Since 2011, metastatic melanoma treatment has patients who experienced toxicity had increased mOS (p <
evolved with commercial approval of BRAF- and MEK-target- 0.001), irrespective of stratification by trial status or therapy.
ed therapy and CTLA-4- and PD-1-blocking antibodies (im- Conclusion: Metastatic melanoma patients receiving com-
mune checkpoint inhibitors, ICI). While novel therapies have mercial treatment may represent a different clinical popula-
demonstrated improved prognosis in clinical trials, few stud- tion with poor prognostic features compared to trial pa-
ies have examined the evolution of prognosis and toxicity of tients. Toxicity may prognosticate treatment benefit.
these drugs among an unselected population. We assess © 2017 S. Karger AG, Basel
whether survival and toxicity reported in trials, which typi-
cally exclude most patients with brain metastases and poor
performance status, are recapitulated within a commercial
access population. Methods: 182 patients diagnosed with C.G. and J.T. share first co-authorship. This work was presented in part
at the Society for Melanoma Research Annual Meeting, San Francis-
stage IV melanoma from July 2006 to December 2013 and co, CA, USA, November 2015 and in a published online abstract for
treated with BRAF- and/or MEK-targeted therapy or ICI were the American Society of Clinical Oncology Annual Meeting, Chicago,
prospectively studied. Outcomes and clinicopathologic dif- IL, USA, June 2016.
132.174.255.215 - 9/5/2017 1:28:47 PM

© 2017 S. Karger AG, Basel Dr. Melissa Ann Wilson

Perlmutter Cancer Center, NYU School of Medicine
160 E. 34th Street, RM 915
UCSF Library & CKM

E-Mail [email protected]
New York, NY 10016 (USA)
www.karger.com/ocl
Downloaded by:

E-Mail Melissa.Wilson @ nyumc.org
 Introduction survival outcomes in an unselected commercial (non-tri-
al) population.
The melanoma treatment landscape was transformed In addition, treatment-related toxicity (adverse events,
in 2010, when randomized controlled trials of novel im- AEs), as reported in clinical trials, may not be broadly ap-
mune and targeted therapies were first to report a statis- plicable to a commercial population because of trial bi-
tically significant survival benefit for metastatic melano- ases in patient selection and specific guidelines for report-
ma [1, 2]. Subsequently, in 2011, ipilimumab, an im- ing AEs. Immunotherapy induces or enhances immune
mune checkpoint inhibitor (ICI) targeting CTLA-4, and responses and triggers immune-related AEs (irAEs) in as
vemurafenib, a BRAF inhibitor, were approved for com- many as 60–85% of patients treated with ipilimumab [2,
mercial use by the US Food and Drug Administration 20]. These irAEs commonly manifest as rash, diarrhea,
(FDA) upon demonstrating improved survival com- colitis, hepatotoxicity, or endocrinopathies [20]. Similar-
pared to standard therapy in phase III studies [1, 2]. ly, targeted therapy may induce AEs via on- or off-target
Pooled analyses and follow-up studies of ipilimumab and effects, manifesting as rash, arthralgias, fatigue, photo-
vemurafenib demonstrated a median overall survival sensitivity, alopecia, nausea, and diarrhea [4]. AEs fre-
(mOS) of 11.4 months [3] and 12 months [4], respec- quently complicate management of patients on targeted
tively, compared with historic mOS of 6–10 months [5] therapy or immunotherapy, and their impact on survival
for patients with distant metastases. Since then, the num- has yet to be fully elucidated.
ber of FDA-approved systemic therapies for melanoma We evaluated survival amongst a more diverse patient
has expanded to include BRAF inhibitor dabrafenib and population treated with commercially available therapies,
MEK inhibitors trametinib and cobimetinib (with com- including patients with characteristics previously exclud-
bination of dabrafenib and trametinib or vemurafenib ed from clinical trials, while simultaneously clarifying the
and cobimetinib considered the standard of care based impact of AEs in treating metastatic melanoma. Specifi-
on phase III trials demonstrating improved OS with cally, we characterized melanoma patients in clinical tri-
BRAF plus MEK inhibition compared to BRAF inhibi- als and commercial cohorts receiving immunotherapies
tion alone), as well as PD-1 inhibitors pembrolizumab, and targeted therapies at Perlmutter Cancer Center,
nivolumab, and the combination of ipilimumab plus New York University (NYU) Langone Medical Center
nivolumab [6–14]. (NYULMC), determined survival and toxicity profiles,
The rapid approval of new drugs and combinations and assessed the impact of toxicity on survival [21–28].
has raised questions about the evolution of prognosis of By offering an extended snapshot spanning clinical trial
metastatic melanoma among the wider patient popula- and commercially accessible periods, we provide a unique,
tion. While survival data from the latest trials are highly comprehensive perspective on the evolution, progress,
encouraging, trial participants may not be representative and critical challenges that define this new era of ad-
of cancer patients in the general population. Clinical tri- vanced melanoma treatment.
als often restrict enrollment of patients with certain co-
morbidities, an Eastern Cooperative Oncology Group
(ECOG) performance status (PS) >1, and active central Methods
nervous system (CNS) disease [15–17]. Furthermore, tri-
Patients
al participants tend to be younger than the general on-
We studied a cohort of adult patients diagnosed with advanced-
cology population [15, 17]. These excluded or under- stage melanoma (defined as unresectable stage III and stage IV
represented groups comprise a significant portion of melanoma) between July 1, 2006 and December 31, 2013 and treat-
melanoma morbidity; approximately 40–50% of stage IV ed at Perlmutter Cancer Center, NYULMC with ipilimumab, pem-
melanoma patients eventually develop clinical manifes- brolizumab, vemurafenib, dabrafenib, or trametinib. All partici-
pants consented to an institutional database for data collection and
tations of melanoma brain metastases (MBM), and 50–
prospective observation maintained on a protocol-driven basis ap-
75% of melanoma patients have MBM at autopsy [18]. It proved by NYU Institutional Review Board (IRB#10362). Treat-
has previously been demonstrated that metastatic mela- ments were excluded if not administered by NYU oncologists, if
noma patients receiving unrestricted access to ipilimu- records omitted pertinent data, if a patient was lost to follow-up,
mab through expanded access programs (EAPs), which or if therapy was administered adjuvantly. Treatments with
nivolumab were not evaluated, as only 5 patients received nivolu-
included patients with ECOG PS >1 and active, untreat-
mab on trial and none received nivolumab as commercial therapy
ed MBM, had worse mOS compared to patients on ear- during the study time frame.
lier clinical trials [19]. Therefore, there is great interest in
132.174.255.215 - 9/5/2017 1:28:47 PM

2 Oncology Goldman  et al.

 

DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 Table 1. Baseline characteristics of clinical trial and commercial therapy cohorts

Clinical trial Commercial p Total

(n = 115; 36.2%) (n = 203; 63.8%) (n = 318)

Age (initial diagnosis)

Mean (SD) 53.6 (14.0) 57.4 (14.9) 0.02 56 (14.7)
Median 54.4 58.9 55.6
Gender
Female 41 (35.7) 63 (31.0) 0.46 104 (32.7)
Male 74 (64.3) 140 (69.0) 214 (67.3)
Primary thickness
Thin 53 (57.6) 68 (44.4) 0.05 121 (49.4)
Thicka 39 (42.4) 85 (55.6) 124 (50.6)
Unknown 23 50 73
Primary ulceration
Absent 41 (51.3) 60 (44.1) 0.33 101 (46.8)
Present 39 (48.8) 76 (55.9) 115 (53.2)
Unknown 35 67 102
Primary site
Head and neck 22 (22.2) 44 (25.0) 0.72 66 (24.0)
Trunk 41 (41.4) 64 (36.4) 105 (38.2)
Extremity 36 (36.4) 68 (38.6) 104 (37.8)
Unknown 16 27 43
BRAF status
Wild-type 37 (37.0) 88 (45.6) 0.17 125 (42.7)
Mutant 63 (63.0) 105 (54.4) 168 (57.3)
Unknown 15 10 25
Age (treatment initiation)
Mean (SD) 59.7 (13.7) 62.4 (14.4) 0.09 61.4 (14.2)
Median 60.2 64.1 62.3
ECOG PS (pretreatment)
0 96 (83.5) 135 (68.2) 0.003 231 (73.8)
>0 19 (16.5) 63 (31.8) 82 (26.2)
Unknown 0 5 5
LDH (stage IV diagnosis)b
Normal 79 (84.9) 129 (75.0) 0.06 208 (78.5)
Elevated 14 (15.1) 43 (25.0) 57 (21.5)
Unknown 22 31 53
Organs with metastases (treatment initiation)
1 16 (13.9) 21 (10.3) 0.20 37 (11.6)
2 or 3 49 (42.6) 107 (52.7) 156 (49.1)
≥4 50 (43.5) 75 (36.9) 125 (39.3)
Metastatic involvement (treatment initiation)c
Visceral only 84 (73.0) 121 (59.6) 0.03 205 (64.5)
CNS only 1 (0.9) 9 (4.4) 10 (3.1)
CNS + visceral 30 (26.1) 73 (36.0) 103 (32.4)
Prior systemic treatment
Yes 77 (67.0) 125 (61.6) 0.40 202 (63.5)
No 38 (33.0) 78 (38.4) 116 (36.5)
Chemotherapy
Yes 48 (41.7) 46 (22.7) <0.001 94 (29.6)
No 67 (58.3) 157 (77.3) 224 (70.4)
Immunotherapy
Yes 34 (29.6) 82 (40.4) 0.07 116 (36.5)
No 81 (70.4) 121 (59.6) 202 (63.5)
Targeted therapy
Yes 19 (16.5) 55 (27.1) 0.04 74 (23.3)
No 96 (83.5) 148 (72.9) 244 (76.7)
132.174.255.215 - 9/5/2017 1:28:47 PM

Outcomes in Metastatic Melanoma Oncology 3

Stratified by Trial vs. Commercial Access DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 Table 1 (continued)

Clinical trial Commercial p Total

(n = 115; 36.2%) (n = 203; 63.8%) (n = 318)

Number of prior treatments

Median (25%, 75%) 1 (0, 2) 1 (0, 2) 0.88 1 (0, 2)
Treatment receivedd
Ipilimumab 60 (52.5) 135 (62.2) <0.001 195 (58.7)
Pembrolizumab 21 (18.3) 12 (5.5) 33 (9.9)
Vemurafenib 25 (21.7) 32 (14.7) 57 (17.2)
Dabrafenib 5 (4.3) 22 (10.1) 27 (8.1)
Trametinib 4 (3.5) 16 (7.4) 20 (6.0)

Figures are n (%) unless indicated otherwise. Percentages are calculated based on known values. SD, standard
deviation; ECOG PS, Eastern Cooperative Oncology Group performance status; LDH, lactate dehydrogenase;
CNS, central nervous system. a Thick primary tumors are defined as ≥2 mm. b LDH >618 U/L is elevated. c Visceral
metastases are metastases to any site except the CNS. d Fourteen patients receiving targeted therapy commercially
were treated with a combination of dabrafenib and trametinib.

Data Collection which were administered in the commercial therapy set-

Differences in survival and toxicity between clinical trial and ting. Fourteen patients received ipilimumab, 10 received
commercial patients receiving immunotherapies and targeted
therapies were assessed. Progression-free survival (PFS) was de- pembrolizumab, and 3 received vemurafenib adminis-
fined as the time from treatment initiation until alternate therapy tered through EAPs and were included within the clinical
was necessitated, disease progression, or death. OS was defined as trial cohort. Baseline patient and disease characteristics of
the time from treatment initiation until death. We also evaluated clinical trial and commercial treatment cohorts are sum-
survival in the context of toxicity and MBM. Collected data in- marized in Table 1. Clinical trial patients had a median
cluded biographical information, primary lesion and disease char-
acteristics, and the treatment course (diagnosis and treatment age of 60.2 years at treatment initiation and 64.3% (n =
dates, trial enrollment, therapies and dosage, AEs, and survival). 74) were male. The majority of clinical trial patients (n =
77, 67.0%) received systemic therapy (chemotherapy, im-
Statistical Analysis munotherapy, or targeted therapy) prior to the treatment
The Fisher exact test was applied to compare demographics, line of interest. Of those clinical trial patients tested, 63%
tumor characteristics, and treatments between trial and commer-
cial cohorts. OS and PFS were assessed by the Kaplan-Meier meth- (n = 63) were BRAF mutant. Patients receiving commer-
od. The p value to compare differences in survival curves was ob- cial treatment had a median age of 64.1 years at treatment
tained from the log-rank test. The Cox proportional hazard mod- initiation and 69.0% (n = 140) were male. The majority of
el for OS was applied to evaluate hazard ratios (HR), 95% confidence commercial treatment patients (n = 125, 61.6%) received
intervals, and p values for univariate and multivariate analysis. Pa- systemic therapy prior to the treatment line of interest. Of
tient treatment lines were considered separately.
those commercial treatment patients tested, 54.4% (n =
105) were BRAF mutant. Commercial treatment patients
were more likely to have an ECOG PS >0 and MBM com-
Results pared to trial patients (31.8 vs.16.5%, p = 0.003, and 40.4
vs. 27%, p = 0.03, respectively) at treatment initiation.
Clinicopathologic and Disease Characteristics While no differences were observed in primary tumor ul-
The 182 patients included in this study were diagnosed ceration or anatomic site, commercial treatment patients
with advanced-stage melanoma between July 1, 2006 and were more likely to have thicker (≥2 mm) primary tu-
December 31, 2013 and subsequently treated with ipilim- mors compared to trial patients (55.6 vs. 42.4%, p = 0.05).
umab, pembrolizumab, vemurafenib, dabrafenib, or tra-
metinib. These 182 patients received 318 discrete treat- Association of Prognostic Features and Survival
ment lines, 115 (36.2%) in the clinical trial setting and 203 We calculated survival stratified by clinical trial and
(63.8%) in the commercial setting. Fourteen patients re- commercial treatment status. No differences in median
ceived therapy with dabrafenib and trametinib, all of PFS were observed between patients receiving therapy
132.174.255.215 - 9/5/2017 1:28:47 PM

4 Oncology Goldman  et al.

 

DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 Color version available online
Immunotherapy and
targeted therapy Immunotherapy Targeted therapy
1.0 1.0 1.0
Trial: n = 81 (mPFS 7.4 m) Trial: n = 34 (mPFS 14.2 m)
Trial: n = 115 (mPFS 8.3 m)
0.8 Commercial: n= 203 (mPFS 5.4 m) 0.8 Commercial: n = 147 (mPFS 5.4 m) 0.8 Commercial: n = 56 (mPFS 5.3 m)

PFS probability

PFS -probability
PFS probability

0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

0 p = 0.13 0 p = 0.15 0 p = 0.94

0 20 40 60 80 0 20 40 60 80 0 10 20 30 40 50 60
a Months b Months c Months

1.0 1.0 1.0

Trial: n = 115 (mOS 17.5 m) Trial: n = 81 (mOS 15.9 m) Trial: n = 34 (mOS 21.4 m)
Commercial: n = 203 Commercial: n = 147 Commercial: n = 56 (mOS 6.6 m)
0.8 0.8 0.8
(mOS 9.2 m) (mOS 12 m)
OS probability

OS probability
OS probability

0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

0 p = 0.0027 0 p = 0.073 0 p = 0.042

0 20 40 60 80 0 20 40 60 80 0 10 20 30 40 50 60 70
d Months e Months f Months

Fig. 1. Kaplan-Meier estimates of progression-free survival (PFS) and overall survival (OS) from treatment ini-
tiation stratified by clinical trial or commercial therapy access. PFS (a) and OS (d) on targeted therapy or immu-
notherapy. PFS (b) and OS (e) on immunotherapy. PFS (c) and OS (f) on targeted therapy. mOS, median OS; m,
months.

commercially and patients on trial when analyzing all ther- 0.29, respectively). For immunotherapies, mOS trended to
apies together (p = 0.13), immunotherapies only (p = 0.15), be lower in patients receiving therapy commercially com-
or targeted therapies only (p = 0.94) (Fig. 1a–c). Patients pared to on trial (12 vs. 15.9 months, p = 0.073) (Fig. 1e).
receiving commercial therapy had significantly lower When analyzing targeted therapies, mOS was significantly
mOS compared to clinical trial patients when assessed lower in patients receiving therapy commercially com-
across all treatment types (9.2 vs. 17.5 months, p = 0.0027) pared to on trial (6.6 vs. 21.4 months, p = 0.0042) (Fig. 1f).
(Fig.  1d). There is no significant difference in 1-year or Cox regression was utilized to assess the prognostic
2-year OS rate comparing the patients receiving commer- value of clinical trial enrollment and clinicopathologic
cial therapy versus on trial when assessed across all treat- characteristics (Tables 2, 3). Treatment with commercial
ments (1-year OS%: 50.8 vs. 63.2%, p = 0.09; 2-year OS%: therapy was independently prognostic for worse OS in
35.6 vs. 43.2%, p = 0.31, respectively). Similarly, there is no univariate (HR = 1.65, p < 0.001) (Table  2) and multi-
significant difference in 1-year or 2-year PFS rate compar- variate analysis (HR = 1.73, p = 0.01) (Table  3). Other
ing the patients receiving commercial therapy versus on variables prognostic for worse OS in univariate analy-
trial when assessed across all treatments (1-year PFS%: sis included thick primary tumors (HR = 1.65, p <
27.3 vs. 36.8%, p = 0.16; 2-year PFS%: 16.7 vs. 23.2%, p = 0.001), ulcerated primary tumors (HR = 1.4, p = 0.04), ele-
132.174.255.215 - 9/5/2017 1:28:47 PM

Outcomes in Metastatic Melanoma Oncology 5

Stratified by Trial vs. Commercial Access DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 Table 2. Univariate analysis to determine factors associated with survival

Variable Baseline Comparison HR (95% CI) p

Trial status Trial Commercial 1.65 (1.24 – 2.20) <0.001

Gender Female Male 1.08 (0.81 – 1.44) 0.60
Primary thickness Thin Thicka 1.65 (1.21 – 2.25) <0.001
Primary ulceration Absent Present 1.40 (1.02 – 1.92) 0.04
ECOG PS (pretreatment) 0 >0 3.93 (2.93 – 5.26) <0.001
LDH (stage IV diagnosis)b Normal Elevated 2.69 (1.96 – 3.69) <0.001
Organs with metastases (treatment initiation) 1 2 0.73 (0.45 – 1.19) 0.21
3 1.14 (0.71 – 1.82) 0.60
≥4 1.54 (1.00 – 2.39) 0.05
Metastatic involvement (treatment initiation) Visceral only CNS + visceral 2.43 (1.84 – 3.23) <0.001
CNS only 2.25 (1.18 – 4.29) 0.01

HR, hazard ratio; ECOG PS, Eastern Cooperative Oncology Group performance status; LDH, lactate dehy-
drogenase; CNS, central nervous system. a Thick primary tumors are defined as ≥2 mm. b LDH >618 U/L is
elevated.

Table 3. Multivariate analysis to determine factors associated with survival

Variable Baseline Comparison HR (95% CI) p

Trial status Trial Commercial 1.73 (1.15 – 2.61) 0.01

Gender Female Male 1.28 (0.83 – 1.96) 0.26
Primary thickness Thin Thicka 1.21 (0.76 – 1.92) 0.43
Primary ulceration Absent Present 1.23 (0.78 – 1.94) 0.36
ECOG PS (pretreatment) 0 >0 4.43 (2.83 – 6.94) <0.001
LDH (stage IV diagnosis)b Normal Elevated 0.84 (0.50 – 1.41) 0.51
Organs with metastases (treatment initiation) 1 2 0.99 (0.38 – 2.53) 0.98
3 0.52 (0.20 – 1.38) 0.19
≥4 1.10 (0.43 – 2.83) 0.84
Metastatic involvement (treatment initiation) Visceral only CNS + visceral 2.83 (1.79 – 4.47) <0.001
CNS only 1.03 (0.26 – 4.06) 0.97

HR, hazard ratio; ECOG PS, Eastern Cooperative Oncology Group performance status; LDH, lactate dehy-
drogenase; CNS, central nervous system. a Thick primary tumors are defined as ≥2 mm. b LDH >618 U/L is
elevated.

vated lactate dehydrogenase (LDH) at stage IV diagnosis Survival Associated with Metastatic Sites
(HR = 2.69, p < 0.001), and metastases to ≥4 organs As MBM have historically been associated with worse
(HR = 1.54, p = 0.05) (Table 2), which were no longer sig- prognosis [18], we assessed survival based on metastat-
nificant in multivariate analysis. ECOG PS >0 was prog- ic sites. Across the entire study cohort, 64.5% (n = 205)
nostic for worse OS in both univariate (HR = 3.93, p < had only visceral metastases, 32.4% (n = 103) had me-
0.001) (Table 2) and multivariate analysis (HR = 4.43, p < tastases of the CNS and viscera, and 3.1% (n = 10) had
0.001) (Table  3). Compared to visceral metastases only only CNS metastases (Table  1). Considering all treat-
(metastases outside the CNS), the presence of both CNS ments together, patients with only visceral metastases
and visceral metastases predicted worse OS in univariate had increased mOS (16.2 months) compared to patients
(HR = 2.43, p < 0.001) and multivariate analysis (HR = with CNS and visceral metastases (6.8 months) and
2.83, p < 0.001), while metastatic involvement limited to CNS only metastases (6.1 months) (p < 0.001) (Fig. 2d),
the CNS predicted worse OS in univariate analysis only with similar results for immunotherapy and targeted
(HR = 2.25, p = 0.01) (Tables 2, 3). therapy individually, the exception being CNS only dis-
132.174.255.215 - 9/5/2017 1:28:47 PM

6 Oncology Goldman  et al.

 

DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 Immunotherapy and
targeted therapy Immunotherapy Targeted therapy
1.0 1.0 1.0 CNS and VISC: n = 33 (mPFS 5.6 m)
CNS and VISC: n = 103 (mPFS 4.5 m) CNS and VISC: n = 70 (mPFS 4.1 m)
CNS only: n = 5 (mPFS 4.1 m)
CNS only: n = 10 (mPFS 4.7 m) CNS only: n=5 (mPFS 7.3 m)
0.8 0.8 0.8 VISC only: n = 52 (mPFS 8.5 m)
VISC only: n = 205 (mPFS 7.4 m) VISC only: n = 153 (mPFS 7.2 m)
PFS probability

PFS probability

PFS probability
0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

p = 0.8 p = 0.73 p = 0.39

0 0 0
0 20 40 60 80 0 20 40 60 80 0 10 20 30 40 50 60
a Months b Months c Months

1.0 CNS and VISC: n = 103 (mOS 6.8 m) 1.0 CNS and VISC: n = 70 (mOS 6.5 m)
1.0 CNS and VISC: n = 33 (mOS 7.5 m)
CNS only: n = 10 (mOS 6.1 m) CNS only: n = 5 (mOS 29.7 m) CNS only: n = 5 (mOS 5.8 m)
0.8 VISC only: n = 205 (mOS 16.2 m) 0.8 VISC only: n = 153 (mOS 17.3 m) 0.8 VISC only: n = 52 (mOS 15.2 m)
OS probability

OS probability

OS probability
0.6 0.6 0.6

0.4 0.4 0.4

0.2 0.2 0.2

p < 0.001 p < 0.001 p = 0.034

0 0 0
0 20 40 60 80 0 20 40 60 80 0 10 20 30 40 50 60 70
d Months e Months f Months

Fig. 2. Kaplan-Meier estimates of progression-free survival (PFS) and overall survival (OS) from treatment ini-
tiation based on sites of metastases (visceral only [VISC], central nervous system only [CNS], or VISC and CNS),
independent of trial enrollment. PFS (a) and OS (d) on targeted therapy or immunotherapy. PFS (b) and OS (e)
on immunotherapy. PFS (c) and OS (f) on targeted therapy. mOS, median OS; m, months.

ease, as there were too few patients for consistent results targeted therapy (Table 4). Severe toxicity was more com-
(Fig. 2e, f). mon with targeted therapy (72.2%) than with immuno-
therapy (29.4%).
Association of Toxicity with Survival Clinical trial and commercial treatment cohorts did
Toxicity was classified according to the presence and not differ in occurrence, degree of overall toxicity, or in
severity (none, mild, or severe). We defined severe toxic- the context of individual organ systems (Table  4). We
ity as grade III or IV AEs or toxicity that necessitated dose evaluated whether survival was impacted by overall toxic-
reductions, temporary interruption (hold events), or per- ity irrespective of trial enrollment. There were no differ-
manent discontinuation of therapy. We defined mild tox- ences in median PFS (Fig. 3a–c). In patients receiving im-
icity as grade I or II AEs and all ungraded, known AEs munotherapy (Fig. 3e), irAEs of any grade conferred in-
that did not cause dose reductions, hold events, or discon- creased mOS compared with no irAEs (p < 0.001): mild
tinuation. Independent of clinical trial enrollment, toxic- irAEs demonstrated the greatest benefit (mOS 16.3
ity was observed in 75.9% (n = 173) of patients receiving months), compared with severe irAEs (mOS 12.2 months)
immunotherapy and 84.4% (n = 76) of patients receiving and no irAEs (mOS 7.2 months). Similarly, in patients
132.174.255.215 - 9/5/2017 1:28:47 PM

Outcomes in Metastatic Melanoma Oncology 7

Stratified by Trial vs. Commercial Access DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 Table 4. Toxicity among patients on immunotherapy and targeted therapy

Trial, n (%) Commercial, n (%) p Total, n (%)

Immunotherapy n = 81 (35.5%) n = 147 (64.5%) n = 228
Toxicity overalla
None 19 (23.5) 36 (24.5) 1.00 55 (24.1)
Mild 38 (46.9) 68 (46.3) 106 (46.5)
Severe 24 (29.6) 43 (29.3) 67 (29.4)
Gastrointestinal
None 51 (63.0) 81 (55.1) 0.40 132 (57.9)
Mild 20 (24.7) 38 (25.9) 58 (25.4)
Severe 10 (12.3) 28 (19.0) 38 (16.7)
Skin
None 44 (54.3) 87 (59.2) 0.69 131 (57.5)
Mild 32 (39.5) 53 (36.1) 85 (37.3)
Severe 5 (6.2) 7 (4.8) 12 (5.3)
Endocrine
None 74 (91.4) 136 (92.5) 0.06 210 (92.1)
Mild 1 (1.2) 8 (5.4) 9 (3.9)
Severe 6 (7.4) 3 (2.0) 9 (3.9)
Nervous system
None 75 (92.6) 131 (89.1) 0.09 206 (90.4)
Mild 1 (1.2) 11 (7.5) 12 (5.3)
Severe 5 (6.2) 5 (3.4) 10 (4.4)
Hold eventsb
0 62 (76.5) 112 (76) 1.00 174 (76.3)
>0 19 (23.5) 35 (24) 54 (23.7)
Targeted therapy n = 34 (37.8%) n = 56 (62.2%) n = 90
Toxicity overalla
None 4 (11.8) 10 (17.9) 0.45 14 (15.6)
Mild 6 (17.6) 5 (8.9) 11 (12.2)
Severe 24 (70.6) 41 (73.2) 65 (72.2)
Gastrointestinal
None 16 (47.1) 29 (51.8) 0.35 45 (50.0)
Mild 7 (20.6) 5 (8.9) 12 (13.3)
Severe 11 (32.4) 22 (39.3) 33 (36.7)
Skin
None 11 (32.4) 23 (41.1) 0.54 34 (37.8)
Mild 13 (38.2) 15 (26.8) 28 (31.1)
Severe 10 (29.4) 18 (32.1) 28 (31.1)
Photosensitivity
None 25 (73.5) 48 (85.7) 0.38 73 (81.1)
Mild 6 (17.6) 5 (8.9) 11 (12.2)
Severe 3 (8.8) 3 (5.4) 6 (6.7)
Musculoskeletal
None 23 (67.6) 36 (64.3) 0.77 59 (65.6)
Mild 6 (17.6) 8 (14.3) 14 (15.6)
Severe 5 (14.7) 12 (21.4) 17 (18.9)
Hold eventsb
0 11 (32.4) 19 (33.9) 1.00 30 (33.3)
>0 23 (67.6) 37 (66.1) 60 (66.7)
Dose reductions
0 23 (67.6) 29 (51.8) 0.19 52 (57.8)
>0 11 (32.4) 27 (48.2) 38 (42.2)

Toxicity was classified according to presence and severity. We defined severe toxicity as grade III or IV AEs
or toxicity that necessitated dose reductions, hold events, or permanent discontinuation of therapy. We defined
mild toxicity as grade I or II AEs and all ungraded, known AEs that did not cause dose reductions, hold events,
or discontinuation. AEs, adverse events; irAEs, immune-related AEs. a Toxicity overall is the highest level of
toxicity across any organ system. b Hold events are temporary interruptions in therapy due to toxicity with intent
to resume therapy after the AE is managed or subsides.
132.174.255.215 - 9/5/2017 1:28:47 PM

8 Oncology Goldman  et al.

 

DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 Color version available online
Immunotherapy and
targeted therapy Immunotherapy Targeted therapy
1.0 1.0 1.0
Mild: n = 11 (mPFS 5.6 m)
Mild: n = 117 (mPFS 6.2 m) Mild: n = 106 (mPFS 6.2 m)
None: n = 14 (mPFS 2.2 m)
0.8 None: n = 69 (mPFS 4 m) 0.8 None: n = 55 (mPFS 4.4 m) 0.8
Severe: n = 65 (mPFS 9.3 m)
Severe: n = 132 (mPFS 7.8 m) Severe: n = 67 (mPFS 6.9 m)

PFS probability

PFS probability
PFS probability

0.6 0.6 0.6

0.4 0.4 0.4

02 02 02

p = 0.015 p = 0.073 p = 0.055

0 0 0
0 20 40 60 80 0 20 40 60 80 0 10 20 30 40 50 60
a Months b Months c Months

1.0 1.0 1.0

Mild: n = 117 (mOS 16.1 m) Mild: n = 106 (mOS 16.3 m) Mild: n = 11 (mOS 7.9 m)
None: n = 69 (mOS 6.2 m) None: n = 55 (mOS 7.2 m) None: n = 14 (mOS 4.5 m)
0.8 0.8 0.8
Severe: n = 132 (mOS 12.3 m) Severe: n = 67 (mOS 12.2 m) Severe: n = 65 (mOS 12.3 m)
OS probability

OS probability

OS probability
0.6 0.6 0.6

0.4 0.4 0.4

02 02 02

p < 0.001 p < 0.001 p = 0.041

0 0 0
0 20 40 60 80 0 20 40 60 80 0 10 20 30 40 50 60 70
d Months e Months f Months

Fig. 3. Kaplan-Meier estimates of progression-free survival (PFS) and overall survival (OS) from treatment ini-
tiation stratified by toxicity. Severe toxicity is defined as grade III or IV adverse events (AEs) or toxicity that ne-
cessitated dose reductions, hold events, or permanent discontinuation of therapy. Mild toxicity is defined as grade
I or II AEs and all ungraded, known AEs that did not cause dose reductions, hold events, or discontinuation. PFS
(a) and OS (d) on targeted therapy or immunotherapy. PFS (b) and OS (e) on immunotherapy. PFS (c) and OS
(f) on targeted therapy. mOS, median OS; m, months.

receiving targeted therapy (Fig. 3f), severe AEs conferred baseline characteristics and survival among clinical trial
increased mOS (12.3 months) compared to mild AEs (7.9 and commercial therapy patients treated with these im-
months) and no AEs (4.5 months) (p = 0.04). munotherapies and targeted therapies. Commercial ther-
apy patients were significantly more likely to have MBM
and worse PS, and they trended to be older at treatment
Discussion initiation and have higher LDH at stage IV diagnosis.
Across all treatment types, treatment on a clinical trial
While phase III clinical trials of BRAF- and/or MEK- compared to the commercial therapy setting was associ-
directed therapies and ICI have demonstrated marked ated with improved OS (17.5 vs. 9.2 months, p = 0.0027),
improvements in survival for metastatic melanoma, there and clinical trial enrollment was independently prognos-
is little systematic data as to how these results translate to tic of survival in univariate and multivariate analysis. We
routine clinical practice. We evaluated differences in suspect differences in baseline patient and disease char-
132.174.255.215 - 9/5/2017 1:28:47 PM

Outcomes in Metastatic Melanoma Oncology 9

Stratified by Trial vs. Commercial Access DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
acteristics are responsible for these observed survival dif- improve survival of patients with MBM. When assessing
ferences. While no differences in PFS were observed, the across all therapies, we observed a significantly reduced
PFS analysis may be confounded by disparate imaging mOS for patients with both CNS and visceral metastases
frequencies for clinical trial and commercial therapy pa- compared to patients with extracranial visceral metasta-
tients, with OS possibly more closely reflecting survival in ses (p < 0.001). A limited number of patients with only
this study. Notably, the two cohorts did not differ in over- CNS metastases were included in this study. Therefore,
all prior systemic treatment for metastatic disease (p = we are unable to draw meaningful conclusions as to
0.40), suggesting survival differences are not a measure of whether these patients have dissimilar survival compared
previous failed treatments and/or end-stage disease. to patients with both CNS and visceral metastases, which
When analyzed by treatment type, clinical trial pa- may reflect different biology of both groups. Nonetheless,
tients treated with targeted therapy demonstrated a sig- the treatment benefits for patients with MBM remain
nificantly increased mOS compared to commercial ther- limited in this commercial therapeutic era. Ongoing
apy patients (p = 0.0042), while clinical trial patients MBM inclusive trials will fully elucidate survival benefits
treated with immunotherapy demonstrated a greater of the most promising combination therapy strategies in
mOS than commercial therapy patients that approached this population.
significance (p = 0.073). Ipilimumab was the first avail- Studies in other cancers have observed more frequent
able ICI, and, thus, it was the most common immuno- and severe toxicity among commercial populations [29],
therapy administered in trial and commercial cohorts. It which may be due to the enrichment of clinical trials with
has previously been demonstrated that patients receiving patients with better PS and fewer comorbidities and, thus,
ipilimumab through an EAP have poorer prognostic fea- less susceptible to AEs [30]. In our study, we found no
tures and worse mOS compared to patients in phase III significant differences in the frequency or severity of AEs
trials [19]. The observation that the survival difference between trial and commercial cohorts when assessing all
between trial and commercial cohorts treated with im- immunotherapies or all targeted therapies. Possibly im-
munotherapy did not achieve significance may reflect the pacting this finding, all trial and commercial patients
inclusion of 24 ipilimumab and pembrolizumab EAP tri- studied were treated at NYU under the supervision of two
al participants, the majority of whom had MBM, in our oncologists, eliminating interinstitutional bias and pro-
study. There was no difference between the number of moting consistent institutional assessment of AEs. Ulti-
prior treatments between the clinical trial and commer- mately, differential toxicity profiles are unable to account
cial cohorts (Table 1). Additional information regarding for observed survival differences between trial and com-
sequencing of treatments in clinical trial and commercial mercial cohorts.
cohorts is included in online supplementary Table 1 (see Although toxicity is never intended, some studies have
www.karger.com/doi/10.1159/000475715 for all online noted the association between off-target immune activa-
suppl. material). tion (irAEs) and ICI melanoma-specific immune re-
These findings of reduced survival in the commercial sponses and its resulting clinical benefit [20]. We investi-
therapy setting have significant implications for trial de- gated the association of irAEs and survival in our cohort,
sign across all cancers, as it is important to translate trial regardless of trial enrollment. Our findings are consistent
findings to clinical practice. Reduced survival in a com- with early ipilimumab trials and a recent study of nivolum-
mercial population has already been reported for other ab that associate irAEs with clinical benefit [31, 32]. In
cancers [29]. For example, Templeton et al. [29] observed one such trial of ipilimumab plus a modified gp100 pep-
a lower mOS for patients with metastatic castration-resis- tide vaccine, only 5% of patients without an irAE of any
tant prostate cancer treated with docetaxel in the com- grade responded, while 36% of patients with severe irAEs
mercial setting compared to on trial (13.6 vs. 20.4 months, clinically responded [33, 34]. Further, patients who expe-
p = 0.007). These distinctions have perhaps more pro- rienced irAEs of any grade while receiving nivolumab on
found clinical implications for melanoma, a cancer with phase I trials showed significant OS differences compared
a markedly diverse natural history and a proclivity to con- to those who did not (p < 0.001) [32]. Our study builds
fer disease characteristics, such as brain metastases, that upon these earlier findings to show that this trend is rep-
have previously excluded patients from trials and confer licated beyond clinical trial populations and may stratify
particularly poor prognosis. according to severity of irAE. While our data suggest that
Treatment of MBM remains a challenge, and our study irAEs may be indicative of treatment benefit, previous
addresses the progress current treatments have made to studies have shown no association between survival, tox-
132.174.255.215 - 9/5/2017 1:28:47 PM

10 Oncology Goldman  et al.

 

DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
icity, or management thereof [35, 36]. Such discrepancies ential toxicity among patients and identify molecular
may result from our broader classification of severe toxic- mechanisms relating AEs and survival remain warranted.
ity, combining trial and commercial cohorts, or pooling Additionally, the fact that Perlmutter Cancer Center,
multiple therapies with a similar mechanism. NYULMC is a single institution tertiary referral center
Interestingly, we also observed an association between could potentially bias the results. However, a subanalysis
toxicity and increased survival in our targeted therapy co- of stage IV melanoma patients diagnosed at Perlmutter
hort, with patients experiencing severe toxicity having the Cancer Center, NYULMC demonstrated consistency
most favorable outcomes. While toxicity is more likely to with the American Joint Committee on Cancer Melano-
be observed in patients on extended courses of therapy, ma Staging Database and another large single institution-
there may be reason to believe that AEs reflect desired, al database, supporting the argument that Perlmutter
on-target MAP kinase pathway inhibition. Studies have Cancer Center, NYULMC’s metastatic melanoma popu-
shown that the occurrence and severity of a rash due to lation reflects the general metastatic melanoma popula-
the targeted therapy erlotinib, an epidermal growth factor tion [40, 41]. Lastly, as patients are often treated with
receptor tyrosine kinase inhibitor utilized for treating multiple subsequent therapies, there might be inflation in
non-small cell lung cancer, is positively associated with statistical significance by analyzing the data by treatment
OS and PFS [37, 38]. In melanoma, vemurafenib-induced lines. However, the treatment lines of the same patient
rash of grade 2 or higher has been associated with early could differ substantially in their outcomes, toxicity, and
2-month tumor response [39]. covariates; thus, we believe there is utility in analyzing the
The observed association between toxicity and in- data by treatment lines.
creased survival is subject to lead time bias and potential In conclusion, we demonstrated that survival reported
confounding, in which patients who live longer or have a in clinical trials may overestimate treatment benefits in a
longer duration of treatments have a higher probability of diverse metastatic melanoma population. Based on these
occurrence of AEs. Therefore, a more rigorous and thor- conclusions, additional verification in large multicenter
ough examination of the association between toxicity and studies is warranted and may provide the proper frame-
survival is warranted. On further review, we assessed the work for obtaining more generalizable outcomes in the
number of treatments for immunotherapy and months new metastatic melanoma treatment era. Additionally,
on treatment for targeted therapy (see online suppl. Ta- we have demonstrated that toxicity from targeted therapy
ble 2). And, although there is no difference in median PFS or immunotherapy may be indicative of clinical benefit.
between patients receiving therapy commercially and pa- Further studies that investigate toxicity as a biomarker for
tients on trial when analyzing all therapies together (p = treatment response are warranted, particularly for target-
0.13), immunotherapies only (p = 0.15), or targeted ther- ed therapy, as this has not been examined sufficiently in
apies only (p = 0.94) (Fig. 1a–c), treatment cycles were the melanoma context.
similar in each toxicity cohort for trial versus commercial
and were similar for no or mild toxicity (median cycle =
4), but interestingly, patients experiencing severe toxicity Acknowledgments
received one less treatment cycle (median cycle = 3).
Funding was provided by the Perlmutter Cancer Center, New
However, there are differences in the median number of
York University Langone Medical Center.
months of treatment between patients on trial and those
commercially treated with target therapy; the median
number of months of treatment was similar in patients Disclosure Statement
with no toxicity and those with severe toxicity (see online
suppl. Table 2). It can be challenging in this analysis to Patrick Ott has received honoraria for advisory board and re-
determine causation, and there is of course, a possible search grants to institution from BMS; has received honoraria for
confounder in lead time bias. advisory board and research grants to institution from Merck; and
has received honoraria for advisory board from Genentech. For
Importantly, there is no evidence that aggressive man- the remaining authors, no conflicts of interest were declared.
agement of AEs diminishes treatment benefit or survival.
This is consistent with results from a study that demon-
strated treatment of irAEs from ipilimumab with system-
ic corticosteroids and anti-tumor necrosis factor did not
diminish survival [36]. Further studies that assess differ-
132.174.255.215 - 9/5/2017 1:28:47 PM

Outcomes in Metastatic Melanoma Oncology 11

Stratified by Trial vs. Commercial Access DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
 References
1 Flaherty KT, Puzanov I, Kim KB, Ribas A, 9 Flaherty KT, Robert C, Hersey P, Nathan P, 16 Kennedy-Martin T, Curtis S, Faries D, Robin-
McArthur GA, Sosman JA, O’Dwyer PJ, Lee Garbe C, Milhem M, Demidov LV, Hassel JC, son S, Johnston J: A literature review on the
RJ, Grippo JF, Nolop K, Chapman PB: Inhibi- Rutkowski P, Mohr P, Dummer R, Trefzer U, representativeness of randomized controlled
tion of mutated, activated BRAF in metastatic Larkin JM, Utikal J, Dreno B, Nyakas M, Mid- trial samples and implications for the external
melanoma. N Engl J Med 2010;363:809–819. dleton MR, Becker JC, Casey M, Sherman LJ, validity of trial results. Trials 2015;16:495.
2 Hodi FS, O’Day SJ, McDermott DF, Weber Wu FS, Ouellet D, Martin AM, Patel K, 17 Kalata P, Martus P, Zettl H, Rodel C, Hohen-
RW, Sosman JA, Haanen JB, Gonzalez R, Schadendorf D: Improved survival with MEK berger W, Raab R, Becker H, Liersch T, Wit-
Robert C, Schadendorf D, Hassel JC, Akerley inhibition in BRAF-mutated melanoma. N tekind C, Sauer R, Fietkau R: Differences be-
W, van den Eertwegh AJ, Lutzky J, Lorigan P, Engl J Med 2012;367:107–114. tween clinical trial participants and patients
Vaubel JM, Linette GP, Hogg D, Ottensmeier 10 Long GV, Stroyakovskiy D, Gogas H, in a population-based registry: the German
CH, Lebbe C, Peschel C, Quirt I, Clark JI, Levchenko E, de Braud F, Larkin J, Garbe C, Rectal Cancer Study versus the Rostock Can-
Wolchok JD, Weber JS, Tian J, Yellin MJ, Jouary T, Hauschild A, Grob JJ, Chiarion-Si- cer Registry. Dis Colon Rectum 2009;52:425–
Nichol GM, Hoos A, Urba WJ: Improved sur- leni V, Lebbe C, Mandala M, Millward M, 437.
vival with ipilimumab in patients with meta- Arance A, Bondarenko I, Haanen JB, Hans- 18 Gorantla V, Kirkwood JM, Tawbi HA: Mela-
static melanoma. N Engl J Med 2010;363:711– son J, Utikal J, Ferraresi V, Kovalenko N, noma brain metastases: an unmet challenge in
723. Mohr P, Probachai V, Schadendorf D, Nathan the era of active therapy. Curr Oncol Rep
3 Schadendorf D, Hodi FS, Robert C, Weber JS, P, Robert C, Ribas A, DeMarini DJ, Irani JG, 2013;15:483–491.
Margolin K, Hamid O, Patt D, Chen TT, Ber- Swann S, Legos JJ, Jin F, Mookerjee B, Fla- 19 Berrocal A, Arance A, Lopez Martin JA, So-
man DM, Wolchok JD: Pooled analysis of herty K: Dabrafenib and trametinib versus riano V, Munoz E, Alonso L, Espinosa E, Lo-
long-term survival data from phase II and dabrafenib and placebo for Val600 BRAF- pez Criado P, Valdivia J, Martin Algarra S:
phase III trials of ipilimumab in unresectable mutant melanoma: a multicentre, double- Ipilimumab for advanced melanoma: experi-
or metastatic melanoma. J Clin Oncol 2015; blind, phase 3 randomised controlled trial. ence from the Spanish Expanded Access Pro-
33:1889–1894. Lancet 2015;386:444–451. gram. Melanoma Res 2014;24:577–583.
4 Larkin J, Del Vecchio M, Ascierto PA, Kra- 11 Robert C, Karaszewska B, Schachter J, Rut- 20 Weber JS, Kahler KC, Hauschild A: Manage-
jsova I, Schachter J, Neyns B, Espinosa E, kowski P, Mackiewicz A, Stroiakovski D, ment of immune-related adverse events and
Garbe C, Sileni VC, Gogas H, Miller WH Jr, Lichinitser M, Dummer R, Grange F, Mortier kinetics of response with ipilimumab. J Clin
Mandala M, Hospers GA, Arance A, Queirolo L, Chiarion-Sileni V, Drucis K, Krajsova I, Oncol 2012;30:2691–2697.
P, Hauschild A, Brown MP, Mitchell L, Ve- Hauschild A, Lorigan P, Wolter P, Long GV, 21 Ribas A, Puzanov I, Dummer R, Schadendorf
ronese L, Blank CU: Vemurafenib in patients Flaherty K, Nathan P, Ribas A, Martin AM, D, Hamid O, Robert C, Hodi FS, Schachter J,
with BRAF(V600) mutated metastatic mela- Sun P, Crist W, Legos J, Rubin SD, Little SM, Pavlick AC, Lewis KD, Cranmer LD, Blank
noma: an open-label, multicentre, safety Schadendorf D: Improved overall survival in CU, O’Day SJ, Ascierto PA, Salama AK, Mar-
study. Lancet Oncol 2014;15:436–444. melanoma with combined dabrafenib and golin KA, Loquai C, Eigentler TK, Gangadhar
5 Garbe C, Eigentler TK, Keilholz U, Hauschild trametinib. N Engl J Med 2015;372:30–39. TC, Carlino MS, Agarwala SS, Moschos SJ,
A, Kirkwood JM: Systematic review of medi- 12 Larkin J, Ascierto PA, Dreno B, Atkinson V, Sosman JA, Goldinger SM, Shapira-Frommer
cal treatment in melanoma: current status and Liszkay G, Maio M, Mandala M, Demidov L, R, Gonzalez R, Kirkwood JM, Wolchok JD,
future prospects. Oncologist 2011;16:5–24. Stroyakovskiy D, Thomas L, de la Cruz-Meri- Eggermont A, Li XN, Zhou W, Zernhelt AM,
6 Hauschild A, Grob JJ, Demidov LV, Jouary T, no L, Dutriaux C, Garbe C, Sovak MA, Chang Lis J, Ebbinghaus S, Kang SP, Daud A: Pem-
Gutzmer R, Millward M, Rutkowski P, Blank I, Choong N, Hack SP, McArthur GA, Ribas brolizumab versus investigator-choice che-
CU, Miller WH Jr, Kaempgen E, Martin-Al- A: Combined vemurafenib and cobimetinib motherapy for ipilimumab-refractory mela-
garra S, Karaszewska B, Mauch C, Chiarion- in BRAF-mutated melanoma. N Engl J Med noma (KEYNOTE-002): a randomised, con-
Sileni V, Martin AM, Swann S, Haney P, Mi- 2014;371:1867–1876. trolled, phase 2 trial. Lancet Oncol 2015; 16:
rakhur B, Guckert ME, Goodman V, Chap- 13 Larkin J, Chiarion-Sileni V, Gonzalez R, Grob 908–918.
man PB: Dabrafenib in BRAF-mutated JJ, Cowey CL, Lao CD, Schadendorf D, Dum- 22 Postow MA, Chesney J, Pavlick AC, Robert C,
metastatic melanoma: a multicentre, open- mer R, Smylie M, Rutkowski P, Ferrucci PF, Grossmann K, McDermott D, Linette GP,
label, phase 3 randomised controlled trial. Hill A, Wagstaff J, Carlino MS, Haanen JB, Meyer N, Giguere JK, Agarwala SS, Shaheen
Lancet 2012;380:358–365. Maio M, Marquez-Rodas I, McArthur GA, M, Ernstoff MS, Minor D, Salama AK, Taylor
7 Robert C, Schachter J, Long GV, Arance A, Ascierto PA, Long GV, Callahan MK, Postow M, Ott PA, Rollin LM, Horak C, Gagnier P,
Grob JJ, Mortier L, Daud A, Carlino MS, Mc- MA, Grossmann K, Sznol M, Dreno B, Wolchok JD, Hodi FS: Nivolumab and ipili-
Neil C, Lotem M, Larkin J, Lorigan P, Neyns Bastholt L, Yang A, Rollin LM, Horak C, Hodi mumab versus ipilimumab in untreated mel-
B, Blank CU, Hamid O, Mateus C, Shapira- FS, Wolchok JD: Combined nivolumab and anoma. N Engl J Med 2015;372:2006–2017.
Frommer R, Kosh M, Zhou H, Ibrahim N, ipilimumab or monotherapy in untreated 23 Ribas A, Gonzalez R, Pavlick A, Hamid O, Ga-
Ebbinghaus S, Ribas A: Pembrolizumab ver- melanoma. N Engl J Med 2015;373:23–34. jewski TF, Daud A, Flaherty L, Logan T,
sus ipilimumab in advanced melanoma. N 14 Wolchok JD, Chiarion-Sileni V, Gonzalez R, Chmielowski B, Lewis K, Kee D, Boasberg P,
Engl J Med 2015;372:2521–2532. Rutkowski P, Grob JJ, Cowey CL, Lao C, Yin M, Chan I, Musib L, Choong N, Puzanov
8 Robert C, Long GV, Brady B, Dutriaux C, Schadendorf D, Ferrucci PF, Smylie M: Up- I, McArthur GA: Combination of vemu-
Maio M, Mortier L, Hassel JC, Rutkowski P, dated results from a phase III trial of nivolum- rafenib and cobimetinib in patients with ad-
McNeil C, Kalinka-Warzocha E, Savage KJ, ab (NIVO) combined with ipilimumab (IPI) vanced BRAF(V600)-mutated melanoma: a
Hernberg MM, Lebbe C, Charles J, Mihal- in treatment-naive patients (pts) with ad- phase 1b study. Lancet Oncol 2014; 15: 954–
cioiu C, Chiarion-Sileni V, Mauch C, Cognet- vanced melanoma (MEL)(CheckMate 067). 965.
ti F, Arance A, Schmidt H, Schadendorf D, J Clin Oncol 2016;34:9505 (suppl abstract).
Gogas H, Lundgren-Eriksson L, Horak C, 15 Lewis JH, Kilgore ML, Goldman DP, Trimble
Sharkey B, Waxman IM, Atkinson V, Asci- EL, Kaplan R, Montello MJ, Housman MG,
erto PA: Nivolumab in previously untreated Escarce JJ: Participation of patients 65 years of
melanoma without BRAF mutation. N Engl J age or older in cancer clinical trials. J Clin On-
Med 2015;372:320–330. col 2003;21:1383–1389.
132.174.255.215 - 9/5/2017 1:28:47 PM

12 Oncology Goldman  et al.

 

DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by:
24 Flaherty L, Hamid O, Linette G, Schuchter L, 30 Seruga B, Sterling L, Wang L, Tannock IF: Re- 36 Horvat TZ, Adel NG, Dang TO, Momtaz P,
Hallmeyer S, Gonzalez R, Cowey CL, Pavlick porting of serious adverse drug reactions of Postow MA, Callahan MK, Carvajal RD,
A, Kudrik F, Curti B, Lawson D, Chapman targeted anticancer agents in pivotal phase III Dickson MA, D’Angelo SP, Woo KM, Pana-
PB, Margolin K, Ribas A, McDermott D, Fla- clinical trials. J Clin Oncol 2011;29:174–185. geas KS, Wolchok JD, Chapman PB: Im-
herty K, Cranmer L, Hodi FS, Day BM, Linke 31 Lutzky J, Wolchok J, Hamid O, Lebbe C, Pe- mune-related adverse events, need for sys-
R, Hainsworth J: A single-arm, open-label, ex- hamberger H, Linette G, de Pril V, Ibrahim R, temic immunosuppression, and effects on
panded access study of vemurafenib in pa- Hoos A, O’Day S: Association between im- survival and time to treatment failure in pa-
tients with metastatic melanoma in the Unit- mune-related adverse events (irAEs) and dis- tients with melanoma treated with ipilimum-
ed States. Cancer J 2014;20:18–24. ease control or overall survival in patients (pts) ab at Memorial Sloan Kettering Cancer Cen-
25 Falchook GS, Long GV, Kurzrock R, Kim KB, with advanced melanoma treated with 10 mg/ ter. J Clin Oncol 2015;33:3193–3198.
Arkenau TH, Brown MP, Hamid O, Infante kg ipilimumab in three phase II clinical trials. 37 Perez-Soler R, Chachoua A, Hammond LA,
JR, Millward M, Pavlick AC, O’Day SJ, Black- J Clin Oncol 2009;27:9034(suppl abstract). Rowinsky EK, Huberman M, Karp D, Rigas J,
man SC, Curtis CM, Lebowitz P, Ma B, Ouel- 32 Freeman-Keller M, Kim Y, Cronin H, Rich- Clark GM, Santabarbara P, Bonomi P: Deter-
let D, Kefford RF: Dabrafenib in patients with ards A, Gibney G, Weber JS: Nivolumab in minants of tumor response and survival with
melanoma, untreated brain metastases, and resected and unresectable metastatic melano- erlotinib in patients with non-small-cell lung
other solid tumours: a phase 1 dose-escalation ma: characteristics of immune-related ad- cancer. J Clin Oncol 2004;22:3238–3247.
trial. Lancet 2012;379:1893–1901. verse events and association with outcomes. 38 Wacker B, Nagrani T, Weinberg J, Witt K,
26 Chapman PB, Hauschild A, Robert C, Haanen Clin Cancer Res 2016;22:886–894. Clark G, Cagnoni PJ: Correlation between de-
JB, Ascierto P, Larkin J, Dummer R, Garbe C, 33 Attia P, Phan GQ, Maker AV, Robinson MR, velopment of rash and efficacy in patients
Testori A, Maio M, Hogg D, Lorigan P, Lebbe Quezado MM, Yang JC, Sherry RM, Topalian treated with the epidermal growth factor re-
C, Jouary T, Schadendorf D, Ribas A, O’Day SL, Kammula US, Royal RE, Restifo NP, Ha- ceptor tyrosine kinase inhibitor erlotinib in
SJ, Sosman JA, Kirkwood JM, Eggermont worth LR, Levy C, Mavroukakis SA, Nichol two large phase III studies. Clin Cancer Res
AM, Dreno B, Nolop K, Li J, Nelson B, Hou J, G, Yellin MJ, Rosenberg SA: Autoimmunity 2007;13:3913–3921.
Lee RJ, Flaherty KT, McArthur GA: Improved correlates with tumor regression in patients 39 Kramkimel N, Thomas-Schoemann A, Gol-
survival with vemurafenib in melanoma with with metastatic melanoma treated with anti- mard JL, Regnier-Rosencher E, Theodore C,
BRAF V600E mutation. N Engl J Med 2011; cytotoxic T-lymphocyte antigen-4. J Clin On- Goldwasser F, Vidal M, Maubec E, Mortier L,
364:2507–2516. col 2005;23:6043–6053. Avril MF, Dupin N, Blanchet B: Severe skin
27 Margolin KA, Hamid O, Weber JS, Pavlick 34 Phan GQ, Yang JC, Sherry RM, Hwu P, Topa- rash during vemurafenib treatment: a predic-
AC, Hodi FS, Amin A, Bennett K, Michener lian SL, Schwartzentruber DJ, Restifo NP, tive factor of early positive response in meta-
T, Minor DR: Ipilimumab retreatment fol- Haworth LR, Seipp CA, Freezer LJ, Morton static melanoma? J Clin Oncol 2014; 32:
lowing induction therapy: the expanded ac- KE, Mavroukakis SA, Duray PH, Steinberg 9092(suppl abstract).
cess program (EAP) experience. J Clin Oncol SM, Allison JP, Davis TA, Rosenberg SA: 40 Balch CM, Gershenwald JE, Soong SJ,
2013;31:9041(suppl abstract). Cancer regression and autoimmunity in- Thompson JF, Atkins MB, Byrd DR, Buzaid
28 Pavlick AC, Chandra S, Stein C, Madden KM, duced by cytotoxic T lymphocyte-associated AC, Cochran AJ, Coit DG, Ding S, Eggermont
Kannan R, Escano C, Muren C, Yepes E, Saba- antigen 4 blockade in patients with metastatic AM, Flaherty KT, Gimotty PA, Kirkwood JM,
do RL, Bhardwaj N: Phase II study of low- melanoma. Proc Natl Acad Sci USA 2003;100: McMasters KM, Mihm MC Jr, Morton DL,
dose cyclophosphamide and ipilimumab in 8372–8377. Ross MI, Sober AJ, Sondak VK: Final version
metastatic melanoma. J Clin Oncol 2014; 35 Ascierto PA, Simeone E, Sileni VC, Pigozzo J, of 2009 AJCC melanoma staging and classifi-
32:e20025(suppl abstract). Maio M, Altomonte M, Del Vecchio M, Di cation. J Clin Oncol 2009;27:6199–6206.
29 Templeton AJ, Vera-Badillo FE, Wang L, At- Guardo L, Marchetti P, Ridolfi R, Cognetti F, 41 Raizer JJ, Hwu WJ, Panageas KS, Wilton A,
talla M, De Gouveia P, Leibowitz-Amit R, Testori A, Bernengo MG, Guida M, Marcon- Baldwin DE, Bailey E, von Althann C, Lamb
Knox JJ, Moore M, Sridhar SS, Joshua AM, cini R, Mandala M, Cimminiello C, Rinaldi G, LA, Alvarado G, Bilsky MH, Gutin PH: Brain
Pond GR, Amir E, Tannock IF: Translating Aglietta M, Queirolo P: Clinical experience and leptomeningeal metastases from cutane-
clinical trials to clinical practice: outcomes of with ipilimumab 3 mg/kg: real-world efficacy ous melanoma: survival outcomes based on
men with metastatic castration resistant pros- and safety data from an expanded access pro- clinical features. Neuro Oncol 2008; 10: 199–
tate cancer treated with docetaxel and predni- gramme cohort. J Transl Med 2014;12:116. 207.
sone in and out of clinical trials. Ann Oncol
2013;24:2972–2977.

132.174.255.215 - 9/5/2017 1:28:47 PM

Outcomes in Metastatic Melanoma Oncology 13

Stratified by Trial vs. Commercial Access DOI: 10.1159/000475715
UCSF Library & CKM
Downloaded by: