Transcript 2006_files/Meropol ASCO 2006 CRC Poster Disc

Understanding the Costs and Benefits of
Colon Cancer Treatment
Colorectal Cancer Poster Discussion 2006
Neal J. Meropol, M.D.
Director, Gastrointestinal Cancer Program
Divisions of Medical Science and Population Science
Fox Chase Cancer Center
“Costs” and “Benefits”
•
•
•
•
•
•
•
Some Costs
Severe toxicity
Symptomatic vs.
laboratory toxicity
Hospitalization
Irreversible toxicity
Quality of life
Death
Dollars, euros, and yen
•
•
•
•
•
•
•
Some Benefits
Response
Progression free
survival
Relapse free survival
Overall survival
Surgical resection
Cure
Quality of life
Deriving a Treatment Valuation
V = Σ (WnpBn) - Σ (WnpCn)
Implications
• There are different types of benefits
• There are different types of costs
• Various stakeholders evaluate costs/benefits
differently
• Individual patients may weigh each cost and
benefit differently, e.g. based upon goals of
therapy, decision making “calculus”
What have we learned today?
The triplet combination of irinotecan,
oxaliplatin and 5FU/LV (FOLFOXIRI) vs
the doublet of irinotecan and 5FU/LV
(FOLFIRI) as first-line treatment of
metastatic colorectal cancer (MCRC):
Results of a randomized phase III trial
by the Gruppo Oncologico Nord Ovest
(G.O.N.O.)
A. Falcone, G. Masi, I. Brunetti, G. Benedetti,
O. Bertetto, V. Picone, S. Chiara, M. Merlano,
S. Vitello, S. Ricci
Falcone et al: Is more better?
Abstract #3513
•
•
•
•
•
•
•
Randomized phase III trial
Unresectable metastatic CRC
Age 18-75
First-line therapy
FOLFIRI vs. FOLFOXIRI
N = 244
Primary endpoint = response rate
FOLFOX vs. FOLFOXIRI: Benefits
FOLFIRI
34%
Response rate
Resection of mets
all (n=244)
6%
liver (n=81)
12%
Median PFS
6.9 mo.
Median S
16.7 mo.
FOLFOXIRI
p
60%
<0.0001
15%
36%
9.8 mo.
22.6 mo.
0.03
0.02
0.0006
0.032
67% in FOLFIRI arm received subsequent FOLFOX
48% in FOLFOXIRI arm received same drugs
subsequently
FOLFIRI vs. FOLFOXIRI: Costs
FOLFIRI
Diarrhea G3-4
12%
Vomiting G3-4
2%
Neurotox G2-3
0%
Neutropenia G3-4
28%
Neutropenia G4
11%
Febrile neutropenia
3%
Toxic death
0%
60-day mortality
1.6%
Hospitalization
?
FOLFOXIRI
p
20%
0.08
7%
0.10
20%
<0.0001
50%
0.0006
17%
5%
0%
1.6%
?
FOLFIRI vs. FOLFOXIRI: Conclusions
• Higher response rate = higher toxicity
• Note: EGFR/VEGF antibodies not routinely
available
• Resection of mets may account for improvements
in long-term outcome
Identify which patients have the potential for longterm survival and CURE with aggressive surgical
management. In this setting, benefits are more
likely to trump costs
Tolerability of
fluoropyrimidines appears to
differ by region
D. G. Haller, J. Cassidy, S. Clarke, D.
Cunningham, E. Van Cutsem, P. Hoff, M.
Rothenberg, L. Saltz, H. J. Schmoll, C.
Twelves
Haller et al: Are there regional variations in
fluoropyrimidine toxicity?
Abstract #3514
• Industry databases from phase III studies,
international enrollment
• Metastatic
– 2 studies, N=1189, capecitabine vs. daily x
5 FU/LV
• Adjuvant
– 1 study, N=1861, CAPOX vs. FU/LV
• Endpoints: adverse events vs. country
Fluoropyrimidine Toxicity (Grade 3-4)
Adjusted Relative Risk (95% CI)
GI Toxicity
Neutropenia
Metastatic
US/RoW
1.72 (1.25-2.36)
1.51 (1.01-2.25)
Adjuvant
US/Asia
RoW/Asia
3.62 (2.11-6.20)
2.38 (1.50-3.77)
0.96 (0.53-1.73)
0.63 (0.41-0.95)
Potential Causes for Regional
Variation in Toxicity
•
•
•
•
•
•
Patient selection (clinical)
Reporting bias
Regional biologic differences
Regional environmental differences
Host-environment interactions
Drug interactions
THF
5-MTHF
Host
MTHFR
DHFR
DHF
Folinic
Acid
5,10-MTHF
Environment
Host-Environment Interactions
Response rate using conventional
criteria is a poor surrogate for
clinical benefit on progression-free
(PFS) and overall survival (OS) in
metastatic colorectal cancer
(mCRC): A comparative analysis of
N9741 and AVF2107
A. Grothey, E. E. Hedrick, R. D. Mass, S.
Sarkar, R. K. Ramanathan, H. Hurwitz, R.
M. Goldberg, D. J. Sargent
Grothey et al: Does benefit
require response?
Abstract #3516
• Retrospective analysis of AVF2107g (N=813)
and N9741 (N=768)
• Response criteria: RECIST (AVF) and WHO
(N9741)
• Hypothesis: Non-responders as well as
responders benefit from the superior therapy
PFS: IFL/Bev vs. IFL
Responders
Non
Responders
PFS: FOLFOX vs. IFL
Responders
Non
Responders
Response Rate is Merely a Surrogate
• Clinical response is not a categorical variable as
RECIST/WHO criteria imply
• Grothey et al analysis supports use of other
endpoints in randomized trials
• Choice of surrogate should depend on clinical
context, goal of study, and treatment mechanism
of action (RR, non-progression rate, PFS, tumor
marker, in vivo pharmacodynamics…..)
• Endpoints must be validated
• Endpoints must permit discrimination of activity
from background; “go – no go” decision
A pooled safety and efficacy
analysis of the FOLFOX4 regimen
(bi-monthly oxaliplatin plus
fluorouracil/leucovorin) in
elderly compared to younger
patients with colorectal cancer
D. J. Sargent, R. M. Goldberg, H. Bleiberg,
A. De Gramont, C. Tournigand, T. Andre, M.
L. Rothenberg, I. M. Tabah-Fisch
Sargent et al: Does age impact efficacy
or toxicity with FOLFOX4?
Abstract #3517
• Pooled efficacy and toxicity analysis of 4 studies
with FOLFOX4 (3 metastatic, 1 adjuvant)
• N=3743, 614 at least 70 years old
• Analyses: age <>70 (3 studies excluded patients
>75 years old)
• Greater G3-4 neutropenia (49 vs. 43%, p=0.04),
thrombocytopenia (5% vs. 2%) in older age group
• No differences in response, survival, dose
intensity, neurotox, GI tox, 60 day mortality (2.3
vs. 1.1%, p=0.2)
Questions Remain…..
• Other key clinical endpoints:
– What was hospitalization rate?
– What was grade 4 neutropenia rate?
– What was dehydration rate?
• What was the age distribution in these studies?
How many patients were old old (e.g. >80)?
• How representative are these study populations
of the general population?
These data highlight the importance of better
characterizing key differences in clinical trial and offstudy populations, and maximizing the
representativeness of clinical trial populations
Direct cost-survival analysis
of therapies for metastatic
colorectal cancer
Y. Wong, N. J. Meropol, D. Sargent, R.
Goldberg, J. R. Beck
Wong et al: What is the costeffectiveness of CRC treatment?
Abstract #3515
• Markov Model comparing cost and
effectiveness of up to 3 lines of treatment to
base cases of 5-FU/LV or FOLFOX
• Cost = total drug costs (average sales price,
70 kg, BSA 1.7 sq.m.)
• Effectiveness = life expectancy
• Aggregate data on progression, survival, and
toxicity from phase II and III published
reports, and N9741 unpublished data
CE Compared to 5-FU/LV
Example: FOLFOX/Bev – Irinotecan -- Cetuximab/Irinotecan
$2700/wk compared to 5FU (survival 117 vs. 55 weeks)
Why does this matter?
• New agents for treatment of metastatic colorectal
cancer result in meaningful improvements in survival
• Cost is largely driven by drug price
• These observations challenge traditional healthcare
cost-effectiveness thresholds
• Cost already limits access to care
• ASCO must work to ensure:
– development of new cancer therapies (innovation)
– access to high quality care for all cancer patients
– appropriate and optimal use of cancer treatments
Expanding Our Understanding of
“Costs” and “Benefits”
• Goals of treatment impact valuation of costs and
benefits
• Biologic and/or environmental factors may influence
treatment toxicity
• Clinical benefit is not restricted to patients who achieve
an objective “response”
• Among clinical trial participants <75 years old, age alone
does not confer substantially greater risk of certain
toxicities with FOLFOX
• Drugs are expensive. We will ultimately need to address
what is an “acceptable” CE ratio, and work with all
stakeholders to ensure access of high quality care for all
patients