Oscar-Della-Pasqua-P..

Report
Personalised Medicine:
Beyond the buzzword
Dr. Oscar Della Pasqua
Clinical Pharmacology
GlaxoSmithKline, United Kingdom
Outline
- Does ‘personalised’ effectively mean the same for clinicians, patients and
industry?
- What are the implications for drug development ?
Effectiveness – integrated measure(s) of efficacy and safety
shift in paradigm from ‘one dose fits all’
shift in paradigm from ‘one endpoint fits all’
shift in paradigm from ‘large, non-enriched trials’
- Model-based approach to integrate data:
right dose, right patient, right methods
- Conclusions
‘Clinical Reality….
We’ve got a new wonder drug!
- But I wonder which
what itdose
will do
to prescribe.
for you.
Treatment Decisions
All patients with same diagnosis
1
Non-responders
or toxic
responders
Treat with alternative drug
2
Responders and patients
not predisposed to toxicity
Treat with
most suitable dose
Biomarkers of Drug Response
Response
Clinical Relevance - Predictive Value
Utility of the information/biomarker
Best
Y
N
Good
Y
Poor
N
N
Variation
Y
N
Y
Y
N
Variation
 Examples
–
–
–
–
Not a biomarker
ErbB-2 over-expression and response to Herceptin
ALOX5 promoter in asthma
CrCL
Bone marrow density
N
Y
CYP2D6 - Polymorphisms
•
Number of functional CYP2D6 alleles (0 - 13) determines
concentrations of nortriptyline. 2 allele patients had
greater clearance than 1 or 0 allele patients.
Conc (nmol/L)
60
•
Lack of efficacy in CYP2D6 x 13 patients
Nortriptyline 25 mg dose
Number of functional
CYP2D6 genes
0
1
30
2
3
13
0
0
24
48
Hours
72
Clinical relevance of CYP2D6
Nortriptyline dosing recommendation in Europe
Clinical Relevance of CYP2D6
Strattera - No Dosing Adjustment
Initial approval 2002, USA
Are the answers to personalised medicine
really here or does one need to look beyond?
(Am J Psychiatry 2002; 159:122–129)
Consider Genetics
Disease / Pharmacokinetics / Pharmacodynamics
C
G
A G C C T A C A T A C T A
A
T
A
C C
T
G
A
C
A
C
G
A
G
C
C
T
G A C C T T C A A T G G A T
A
G
A
A
T
T C A A A G T A
G
A
T A C G A T G
A
A T G G A A T
A
G
A
T T A A G G T
T
A
C
C
A
T
A A G C C
T
A
A
G
G
T
A
T
A
T
A A C C T A T
T
G
C
A G C A A T A G T
Epidemiology / Genetics / Clinical Pharmacology
Can one predict the impact of variability or
noise in drug effect with a single marker?
What do you see when you have spent 8 months designing a sports car?
Consider Intrinsic and Extrinsic Factors
Disease / Pharmacokinetics / Pharmacodynamics
C O M E D
P
R
O
T
E
I
G E N E
B
I
N
D I
I
N
T A R G E
D
I C A T I O N S
E
T
A
B O D
R
Y
T
F
R
A
A
C
N
T
S
T I C P O L Y M O R P
R
O
S
R
T
S E A S E S
E
R
S
T S
H E
Y W
H I
E
D
R
U
G
M
E
T
A
B
O
L
I
S
I
N
G
B A L S
I G H T
M S
Z Y M E S
E C E P T O R S
Epidemiology / Genetics / Clinical Pharmacology
Model-based Approaches for Prediction of Response
Disease / Pharmacokinetics / Pharmacodynamics
C O M E D
P
R
O
T
E
I
G E N E
B
I
N
D I
I
N
T A R G E
D
I C A T I O N S
E
T
A
B O D
R
Y
T
F
R
A
A
C
N
T
S
T I C P O L Y M O R P
R
O
S
R
T
S E A S E S
E
R
S
T S
H E
Y W
H I
E
D
R
U
G
M
E
T
A
B
O
L
I
S
I
N
G
B A L S
I G H T
M S
Z Y M E S
E C E P T O R S
Epidemiology / Genetics / Clinical Pharmacology
BeSt study design
•
•
Retrospective, multi-centre, open
509 patients with active RA enrolled in this study are
participants in a trial to test the effectiveness of
different treatment strategies (BeSt- study)
•
all patients have active disease according to ACR
criteria, disease duration < 2 years
•
247 patients are treated with monotherapy MTX
Wessels et al. Arthritis Rheum.56:1765-75, 2007
BeST study: summary
DAS >2.4
205 RA patients
Active RA at baseline DAS 4.5
DAS  2.4
MTX 15 mg/week or 25 mg/week,
folic acid 1 mg/day
RESPONSE 47% at 6 months
ADVERSE DRUG EVENTS 30%
age
gender
hormonal status
co-morbidity
ethnicity
previous DMARD use
HLA-DRB1 alleles (shared epitope)
PTPN22
Cytochrome P450 enzymes
Candidate genes
Genes
Disease duration
disease activity
Anti-CCP
rheumatoid factor
Host
Disease
Life style e.g. smoking and diet
social class
Environment
Drug
Treatment outcome
Factors influencing
outcome
Disease activity score
ACR criteria
Health assessment questionnaire
Radiographic score
Measures to
evaluate outcome
Folate pathway
RFC
MTX response
MTHFR haplotype as factor for
100
88
80
77
66
62
46
46
50
good clinical
response
MTHFR testing may
determine which RA
patients will benefit
from MTX
good clinical
improvement
ie
2
co
p
ie
s
co
p
1
0
co
p
ie
s
moderate clincal
imrpovement
Number of MTHFR 677C1298A haplotype copies
genetics contribute to MTX treatment outcome in RA
‘Adenosine release’
Good clinical response with MTX
at 6 months (%)
AMPD1 T-allele, ATIC CC
genotype, ITPA CC
genotype are 2-3 fold more
likely to achieve good
clinical response
93
50
58
60
61
68
75
37
26
42
37
41
ITPA
43
43
AMPD
47
Current MTX pharmacogenetic research
From associations
with genes
to a predictive clinical
tool
“MTX sensitive RA”
- Simple model
- validation in 2nd cohort
Development of a predictive model of
clinical response
24 baseline variables believed
to influence RA disease state
and MTX drug response were
selected based on literature
RFC
17 SNPs in 13 genes
involved in the MTX
mechanism of action,
purine and pyrimidine
synthesis
ITPA
AMPD
Factors determining efficacy for
individual MTX monotherapy
Baseline Variable
Gender
Score
Female
premenopausal
postmenopausal
Male
Disease activity
DAS at baseline 3.8
≤ but
DAS at baseline >3.8,
DAS at baseline >5.1
5.1
≤
1
1
0
0
3
3.5
Immunological factors
Rheumatoid factor negative and non-smoker
Rheumatoid factor negative and smoker
Rheumatoid factor positive and non-smoker
Rheumatoid factor positive and smoker
0
1
1
2
Genetic factors
MTHFD1 1958 AA genotype
AMPD1 34 CC genotype
ITPA 94 A- allele carrier
ATIC 347 G-allele carrier
Other genotypes
1
1
2
1
0
Suggestions for clinical application
of the model
Categories
Clinical consequence
Scores ≥ 6
Low probability to respond to MTX
monotherapy. Consider a combination strategy.
Scores < 6, but > 3.5
Intermediate probability to respond to MTX
monotherapy. Evaluate after 3 months therapy.
Scores ≤ 3.5
High probability to respond to MTX
monotherapy Dose escalation to 25 mg/week
if necessary.
Receiver Operating Curves (ROC)
1,0
0,8
sensitivity
0,5
PG Model:
pharmacogenetic model
True positive response 95%
0,3
(36 out of 38)
non-genetic model
Non-genetic model
True negative response
87% (62 out of 72)
Percentage of patients
categorised: 32%
Percentage of patients
categorized: 60%
0,0
0,3
0,5
1- specificity
0,8
1,0
Conclusions - BeST
The chance to achieve clinical response with MTX
treatment is predictable in recent onset RA.
It is feasible to assist initial treatment decisions
to tailor therapy in RA patients according to their baseline criteria
(symptoms, signs and genotype)
Model-based Approaches for Dose Optimisation
Disease / Pharmacokinetics / Pharmacodynamics
C O M E D
P
R
O
T
E
I
G E N E
B
I
N
D I
I
N
T A R G E
D
I C A T I O N S
E
T
A
B O D
R
Y
T
F
R
A
A
C
N
T
S
T I C P O L Y M O R P
R
O
S
R
T
S E A S E S
E
R
S
T S
H E
Y W
H I
E
D
R
U
G
M
E
T
A
B
O
L
I
S
I
N
G
B A L S
I G H T
M S
Z Y M E S
E C E P T O R S
Epidemiology / Genetics / Clinical Pharmacology
New Technologies – Old tools?
From Biomarker data to Treatment Decision
JAMA, 296 (12), 2006
The concentration-response surface:
Efficacy
What is the surface for a given population /patient group?
Where are you during development?
Multidimensional Diseases
- Multiple Endpoints 1.
Migraine (4)
2.
Alzheimers (2)
3.
Acute Pain (3)
4.
Lower Back Pain (3)
5.
Sleep Disorders (3 or 6)
6.
RA (4)
7.
OA for symptom modif. (2)
8.
Asthma, COPD (2)
9.
ED (3)
10.
Skin Aging (2)
11.
Menopausal Symptoms (3)
19.
Organ Transplantation (2)
12.
Fracture Healing (2)
20.
Primary Biliary Cirrhosis (4)
13.
Acne (4)
21.
BPH (2)
14.
Male Pattern Baldness (2)
22.
Multiple Sclerosis (2)
15.
Glaucoma (9)
23.
Epilepsy (3)
16.
17.
Ophthalmology – dry eye (2) 24.
25.
Hepatitis B (up to 3)
18.
Vaginal Atrophy (3)
26.
Vaccines (up to 23)
Operable Breast Cancer
(with + auxiliary lymph nodes) (2)
Fibromyalgia (2-3)
Model-based risk assessment
Model-based risk assessment
Model-based Approaches:
Dosage strategy for enoxaparin
Observed vs. population predicted anti-Xa concentrations for the
two-compartment model with CrCL and weight covariates in the
model. Individual data points are shown as dots and the unity as
a solid line
Three-dimensional surface showing the relationship between
CrCL, weight and predicted Css. The surface shows how the Css
changes with both weight and CrCL simultaneously
Feng et al (2007), Br J Clin Pharmacol 62:165–176
general medical unit
8.3 IU/ kh/h
5.8 IU/kg/h
5.0 IU/kg/h
4.2 IU/kg/h
% Css out of range
% Css >1.2 UI/ml % Css < 0.5 UI/ml
intensive care unit
(1, CrCL <30 ml min−1; 2, CrCL 30–50 ml min−1; 3, CrCL >50 ml min−1).
Model-based Dose Recommendations
Barras et al. (2007) Clin Pharmacol Ther advance online publication doi:10.1038/sj.clpt.6100399
Sotalol in SVT
Sotalol conc (ug/mL)
Probability of arrhythmia suppression in the 15 children
with supraventricular tachycardia vs sotalol trough
concentration under steady-state conditions and an 8-h
dosing interval.
Filled circles 6 neonates (28 days).
Effect of Age on Clearance
Sotalol oral Clearance
(ml/min/kg)
Probability of Response
PK/PD relationship
Age (years)
Measured (closed diamonds) and model predicted oral
sotalol clearance based on body weight (open
diamonds). Median (solid line) and the 10th and 90th
percentile (dashed line) of 1,000 simulated data sets.
Dose
Recommendation
Age-specific Dosing regimen for sotalol in
children with SVT
Black box plots and hatched bars indicate recommended dosing range. (A) Simulated sotalol trough concentrations (125 patients
per group and dose level) for paediatric patients with supraventricular tachycardia. Lines indicate 50% and more than 95% efficacy.
(B) Patient fraction with 50% and more than 95% probability of arrhythmia suppression. Arrows
indicate start and target doses.
Summary
- Does ‘personalised’ effectively mean the same for clinicians, patients and
industry?
- What are the implications for drug development ?
Effectiveness – integrated measure(s) of efficacy and safety
shift in paradigm from ‘one dose fits all’
shift in paradigm from ‘one endpoint fits all’
shift in paradigm from ‘large, non-enriched trials’
- Model-based approach to integrate data:
right dose, right patient, right methods
- Conclusions
Personalised Treatment: Delicate Balance Between
Benefit and Risk
The greatest obstacle to discovery is not
ignorance, but the illusion of knowledge
by Daniel Boorstin

similar documents