Intermittent versus Continuous Systemic
Therapy for Metastatic Colorectal Cancer
PRO: Continue Systemic Therapy
Deb Schrag, MD, MPH
Presentation in “Great Debates” Symposium
March 28th, 2014
Attending Physician, Dana-Farber Cancer Institute
Professor of Medicine, Harvard Medical School
Boston MA, USA
Overview
• Brief review of stop versus go trials
– COIN
– OPTIMOX1 and 2
– DREAM
– CAIRO-3
• Challenges to interpretation of these studies
• Strategies for clinical decision making
Common Clinical Decisions
• What is optimal management of mCRC post-induction
chemo?
• How long to continue induction systemic rx?
• Should maintenance therapy:
–
–
–
–
–
Be identical to induction?
Drop oxaliplatin?
Include 5FU/Bev?
Bev alone, Bev/erlotinib?
No maintenance?
• How long is it reasonable to continue chemo breaks?
Stop and Go Strategies for PostInduction Systemic Therapy Are Not
All Created Equal
Strategy for post-induction
systemic treatment for mCRC
patients who do not have PD
Plan to alternate intense/light
phases of treatment..eg 2 months
on 2 months off systematically
Eliminate the most toxic ingredient
eg oxaliplatin or irinotecan but
restart the more intense regimen at
progression
Eliminate cytotoxics and continue
biologics
Eliminate all therapy….true “holiday”
Example trial that uses this
strategy
GISCAD
OPTIMOX-1
DREAM/OPTIMOX-3, SAKK
OPTIMOX 2, CAIRO-3
Stop and Go Trials’ Heterogeneity
Makes them Challenging to Compare
• Different induction regimens/durations
• Different maintenance regimens/durations
• Various endpoints
•
•
•
1st PFS interval, 2nd PFS interval
Duration of disease control
OS
• Monitoring strategy influences PFS assessment
• CEA frequency
• Scan frequency
Continuous v Intermittent Therapy:
The MRC Coin Trial
Responding
or stable
disease
after 12
weeks
Maughan et al The
Lancet. 2003. 361:
457-64.
Continuous v Intermittent Therapy:
MRC COIN Trial
•
•
•
•
•
CAPOX or FOLFOX until PD vs. CAPOX/FOLFOX for 12 weeks with reinitiation of same chemo at progression for another 12 weeks
Median off-treatment duration with intermittent therapy was 4.3 months
Significantly fewer adverse events
Overall survival was similar in both groups
Intermittent strategy didn’t meet non-inferiority threshold
PFS HR
1.20 (0.96-1.49)
favors continuous
Maughan et al The Lancet. 2003. 361: 457-64.
OPTIMOX Studies
FOLFOX4 until TF
OPTIMOX-1
N = 620
FOLFOX7
FOLFOX7
sLV5-FU2
Tournigand et al, JCO 2006
mFOLFOX7
OPTIMOX-2
N = 202
mFOLFOX7
sLV5-FU2
mFOLFOX7
mFOLFOX7
CFI
Maindrault-Goebel et al, ASCO 2007 Abstract #4013
OPTIMOX 1:
Hold the oxaliplatin post-induction
R
A
N
D
O
M
I
Z
e
FOLFOX4 until
Progression
FOLFOX7 x 6 cycles
sLV5FU2 x up to 12 cycles
FOLFOX7 x 6 cycles
N=311
N=309
Only 40%
reintroduced oxaliplatin
18.5% early progression/death
18.4% toxicity (including neuropathy)
5.5% surgery
17.5% unknown
Tournigand et al. JCO 2006;24:394-400
OPTIMOX 1: Maintenance 5FU Alone
No meaningful benefit is clearly evident from
continuation of oxaliplatin without a break
Caveat: most of us don’t use FOLFOX4 or 7
PFS
OS
Tournigand et al. JCO 2006;24:394-400
OPTIMOX 1: Maintenance 5FU alone
Grade 3 / 4 Toxicity
Continuous
therapy has:
• higher overall
toxicity
• substantially
higher late
neurotoxicity
Grade 3 Neurotoxicity
Tournigand et al. JCO 2006;24:394-400
OPTIMOX 2: Go and Full Stop
Chibaudel B et al. JCO 2009;27:5727-5733
OPTIMOX 2: Go and Full Stop
HR= 0.71 (95% CI, 0.51 to 0.99;
P = .046
Median duration of maintenance therapy = 4.8 months in the arm 1
Median duration of chemotherapy free interval = 3.9 months in arm 2.
Chibaudel B et al. JCO 2009;27:5727-5733
OPTIMOX 2: Go and Full Stop
A chemotherapy holiday with no maintenance therapy has significantly worse
PFS compared to maintenance therapy (allowing for attenuating oxaliplatin).
Overall survival trend also favors maintenance therapy but is not significant
PFS
HR = 0.61; P = .0017
OS
HR = 0.88; P = 0.42
Chibaudel B et al. JCO 2009;27:5727-5733
MACRO Trial
Capecitabine
Oxaliplatin
Bevacizumab
x6 cycles q3w
Capecitabine
Oxaliplatin
Bevacizumab
until progression
Progression
R
N=480
Capecitabine
Oxaliplatin
Bevacizumab
x6 cycles q3w
Bevacizumab
until progression
Non-inferiority design
Assuming 10 months as median PFS
Non-inferiority limit of 7.6 m (HR=1.32)
One sided alpha = 0.025, one side; Power = 80%
Taberno et al ASCO 2010
MACRO Trial:
Non-significant PFS trend favors continuation of XELOX
LNI: 1.32
Patients at risk
Taberno et al ASCO 2010
MACRO Trial
No difference in overall survival
Patients at risk
Taberno et al ASCO 2010
The GERCOR DREAM phase III trial:
Bevacizumab with or without erlotinib maintenance
following induction 1st-line chemotherapy plus
bevacizumab, in patients with metastatic CRC
CC. Tournigand et al ASCO 2012
• 6-12 cycles of Induction Chemotherapy
– EITHER FOLFOX-7 or XELOX or FOLFIRI all with Bev
• If no Disease Progression then randomize 1:1 to
• Maintenance Bev and Erlotinib
or
• Maintenance Bev alone
DREAM Results
Bev vs. Bev/Erlotinib maintenance
B
222
B+E
223
HR
P value
PFS from
randomization
4.6
5.8
0.73
P=.005
PFS from
registration
9.2
10.2
0.75
P=.005
Grade 3 skin
toxicity
0%
20%
-
-
Overall survival no difference: 25.4 months [95% CI 22.9–28.2]
TAKE HOME: Marginal survival benefit and excess toxicity for
Bev/Erlotinib
SAKK: Induction +/- Maintenance Bev
BEV continuation
(7.5 mg/kg q 3 w)
until PD
First-line chemotherapy + BEV
for 4-6 months
Stratification factors:
No
PD
Randomization
1: 1
No antitumor treatment
(no BEV) until PD
• Best response during first-line chemotherapy + BEV (CR/PR vs SD)
• Duration of first-line chemotherapy + BEV (16-20 vs 21-24 weeks)
• Type of chemotherapy (Irinotecan + 5-FU vs Oxalipaltin + 5-FU vs
Fluoropyrimidine mono)
• Disease burden (metastases in one organ vs multiple organs)
Study conducted in 26 sites in Switzerland (accrual period 2007-2012)
SAKK: Induction +/- Maintenance-B
TTP
(from randomization)
P ro p o rtio n w ith o u t p ro g re ss io n
1.0
BEV
no BEV
0.8
/
BEV
no BEV
No. of events
124
123
Median
(95%CI)
4.1 months
(3.1-5.4)
2.9 months
(2.8-3.8)
/
0.6
/
HR
95% CI
0.4
TTP Trend
favors
continued Bev
0.74
(0.57-0.95)
Non-inferiority
p = 0.47
0.2
/
//
/
/
/
/
/
/
0.0
0
6
12
18
No. at risk
24
30
36
42
48
Time (months)
BEV
131
40
14
8
6
5
3
2
1
no BEV
131
22
10
7
5
1
1
1
0
Koeberle ASCO 2013
/
SAKK: Induction +/- Maintenance-B
P ro p o rtio n w ith o u t p ro g re ss io n /d e a th
PFS
(from start of first-line therapy)
1.0
BEV
no BEV
0.8
0.6
/
BEV
no BEV
No. of events
125
124
Median
(95%CI)
9.5 months
(8.6.-10.2)
8.5 months
(8-8.9)
HR
95% CI
0.4
0.75
(0.58-0.96)
Difference
PFS
Trend favors
continued Bev
p = 0.021
/
0.2
/
/
/ /
/
/ /
/
/
/
/
0.0
0
6
12
18
No. at risk
24
30
36
42
48
54
Time (months)
BEV
131
122
40
13
6
6
5
3
2
1
no BEV
131
116
18
8
7
4
1
1
0
0
Koeberle ASCO 2013
SAKK: Induction +/- Maintenance-B
Overall Survival
(from start of first-line therapy)
1.0
/// /
BEV
/
//
/
no BEV
//
/
/
/
P ro p o rtio n s u rv ivin g
0.8
//
/
///
BEV
no BEV
No. of events
84
84
Median
(95%CI)
25.1 months
(22-28.9)
22.8 months
(20.3-26.1)
/
/ /
0.6
/
/
/
// //
/
///
//
/
/ //
///
/ /
0.4
HR
95% CI
/
/
/
/
0.83
(0.61-1.12)
/
/
/ / // /
//
Difference
/
/
p = 0.218
//
/ / / / // / / /
/
0.2
/
// /
/
//
/
/
/
0.0
0
6
12
18
No. at risk
24
30
36
42
48
54
60
Time (months)
BEV
131
130
115
86
52
33
22
10
3
1
0
no BEV
131
131
107
76
44
25
13
6
1
1
1
Koeberle ASCO 2013
OS Trend
favors
continued Bev
CAIRO-3: Study design
PFS1
PFS2
observation
SD/PR/CR post
CAPOX-B x 6
PD
R
Re-introduction
CAPOX-B
PD
capecitabine +
bevacizumab
Primary endpoint: PFS2
• time from randomization to progression upon re-introduction of CAPOX- B
• PFS2 is considered to be equal to PFS1 for patients in whom CAPOX- B is not
reintroduced after PFS1 for any reason
Koopman: ASCO 2013 for Dutch Colorectal Study Group
CAIRO-3: Study design
PFS1
TT2PD
observation
SD/PR/CR post
CAPOX-B x 6
PD
R
Any further Rx
Including CAPOX-B
capecitabine +
bevacizumab
•
•
TT2PD = time to second progression of disease
= time from randomization to progression after any post PFS1 Rx
including CAPOX-B
Koopman: ASCO 2013 for Dutch Colorectal Study Group
PD
CAIRO-3
Primary endpoint PFS2
1.0
median PFS2 - Observation : 10.5 (95% CI: 9.3 - 12.3 )
median PFS2 - Maintenance : 11.8 (95% CI: 10.2 - 13.3 )
Median PFS2
0.8
ITT, events/n ( 246 / 279 - 243 / 279 )
Observation
10.5 m [95%CI: 9.3-12.3]
HR= 0.81 ( 95% CI: 0.67 - 0.98 )
Maintenance
11.8 m [95%CI: 10.2-13.3]
PFS2 Probability
stratified log-rank p-value 0.028
0.6
Stratified HR
0.81
p value
0.028
[95%CI: 0.67-0.98]
adjusted HR 0.77, p 0.007
0.4
Maintenance
0.2
Observation
0.0
279
207
111
42
16
11
4 Observation
279
207
130
66
38
23
12 Maintenance
0
6
12
18
Time (mths)
24
30
36
CAIRO-3 TT2PD
1.0
median TT2PD - Observation : 15.0 (95% CI: 13.6 - 16.4 )
median TT2PD - Maintenance : 19.8 (95% CI: 18.0 - 21.9
)
Median
Observation
ITT, events/n
( 223 / 279 - 251 / 15.0
279 )
0.8
Maintenance
TT2PD
m [95%CI:13.6-16.4]
19.8 m [95%CI: 18.0-21.9]
TT2PD Probability
HR= 0.67 ( 95% CI: 0.55 - 0.81 )
Stratified
HR p-value
0.67
stratified log-rank
0
p value
0.6
[95%CI: 0.55-0.81]
< 0.00001
Maintenance
adjusted HR 0.63, p <0.001
0.4
Observation
0.2
0.0
279
247
174
97
52
36
13 Observation
279
251
187
134
87
52
31 Maintenance
0
6
12
18
Time (mths)
24
30
36
CAIRO-3 Overall Survival
1.0
median OS - Observation : 18.2 (95% CI: 16.3 - 20.8
)
Median
OS
median OS - Maintenance : 21.7 (95% CI: 19.4 - 24.0 )
Observation
0.8
18.2 m [95%CI: 16.3-20.8]
ITT, events/n ( 204 / 279 - 217 / 279 )
Maintenance
21.7 m [95%CI: 19.4-24.0]
Stratified
0.87
HR=
0.87 ( 95% CI:HR
0.71 - 1.06
)
[95%CI: 0.71-1.06]
OS Probability
stratified
log-rank p-value 0.156
p value
0.156
0.6
Maintenance
adjusted HR 0.80, p 0.035
Observation
preliminary survival analysis
0.4
0.2
0.0
279
248
184
122
78
53
28 Observation
279
252
192
143
95
58
33 Maintenance
0
6
12
18
Time (mths)
24
30
36
Cancer Care Ontario
Metanalysis
•
•
•
•
Includes 10 trials, but not CAIRO3
Notes heterogeneity of stop/go strategies
Variation in maintenance regimens
Inconclusive as to benefits
Evidence Summary from Trials
• No strategy is clearly superior
• The preponderance of evidence favors some
form of maintenance therapy
• Reasonable to drop the oxaliplatin
(OPTIMOX-1)—low threshold to do so
• Reasonable to continue 5FU-B (CAIRO)
Gompertzian Model of Tumor Growth
and Norton–Simon hypothesis
The Norton-Simon hypothesis states
that the rate of cancer cell death in
response to treatment is directly
proportional to the tumor growth rate at
the time of treatment.
Schmidt C JNCI J Natl Cancer Inst 2004;96:1492-1493
Factors to Consider in Counseling mCRC
patients about “Chemo Holidays”
Tilt towards maintenance:
• Continued response could result
in candidacy for R0 resection
• Major/brisk/ongoing response
• More symptoms from cancer
than from chemotherapy
• Easy to track disease status—
CEA or measurable disease
• mCRC is primary threat to
survival
• Tolerates chemotherapy well
• Prefers intensive management
Tilt towards true holiday
• Will never be a candidate for
R0 resection
• Stable disease/slow response
• Minimal disease burden
• Tolerates therapy poorly
• Major comorbidity
• Prefers less
intensive/interventionist
management
Viewpoint
• We are saturated on maintenance regimen trials
• Time to tailor strategies to disease biology
• If biology doesn’t provide a clear cut signal-- patient
preference is the key issue
• Better understanding of resistance mechanisms and
genomics should enable us to distinguish low vs. high
grade tumors with greater accuracy
Are further trials of continuous versus
intermittent treatment important?
• With current agents, seems unlikely that
additional trials will provide greater
clarity
• May yield some insight from subgroup
analyses based on genomic
charachteristics of mCRC
• genomic analyses of olecular subtypes
by continuous vs. intermittent strategy
neede