Assay Improvement for Protein Therapeutics at Clinical Development Stage
AMV final 20Jun17
1. Stephan O. Krause, PhD
Director, QA Technology
AstraZeneca Biologics
PDA Biosimilar Conference
(co-sponsored by PQRI)
20-21 June 2016
Analytical Method Validation for
Biosimilars
2. 2
Outline
1) Review of strategic opportunities to reduce analytical
method lifecycle steps for accelerated programs
- Analytical platform technology (APT) methods for Tier 2 category
- Product and process characterization APT methods for Tier 3
category
- APT conditions
2) Review of need for increased analytical method performance
- Difference(s) and challenges for Biosimilars
- AMV acceptance criteria for Tier 1 and 2
2
3. The content and views expressed in this presentation
by the author are not necessarily views of the
organization he represents.
3
4. 4
Analytical Platform Technology (APT) Method
• An APT method is an analytical method used for multiple products
and/or types of sample matrix without modification of the
procedure.
• Similar to compendial methods, an APT method may not require full
validation for each new product or sample type.
• A test method becomes an “approved” (APT) method when
included in a marketing license application, and the license is
approved.
4
5. 5
Typical CQA Development, CMC Changes, and Specifications
From: Krause, S., WCBP, 30Jan13, Washington, DC.
FTIH POC BLA
Tox Studies
Phase 1
Phase 2
Phase 3
Clinical Resupply
Mfg/Formulation Change(s)
Specifications
Revision(s)
Negotiations, Final
Commercial Specifications
QTPP
Final CQAs &
Control Strategy Approval
Potential CQAs
Product & Process
Design
Life-Cycle
Management
POST-APPROVAL
CHANGES
PHASE 3PHASE 1/2Pre-IND
CQADevelopment
(QbDProcess)
SpecsLifeCycle
Mgmt
CMCandTech
TransferProcess
Analytical
Manufacturing
Strategic or
Tactical Changes
Method
qualification
Dose
change
Delivery
Device
PQ lots
Setting of Initial
Specifications
Specifications
Revision(s)
Mfg
Transfer
Method
validation
Method
transfer
Formulation
Change
Process Verification
Method Maintenance
Global
Supply
Commercial
Specifications
6. Accelerated CQA Development, CMC Changes, and Specifications
6
FTIH POC BLA
Tox Studies
Phase 1
Phase 3
Clinical Resupply
Mfg/Formulation Change(s)
Specifications
Revision(s)
Commercial
Specifications
Negotiations, Final
Commercial Specifications
and/or Post-BLA
commitmens
QTPP
Final CQAs &
Control Strategy Approval
Potential CQAs
Product & Process
Design
Life-Cycle
Management
POST-APPROVAL
CHANGES
PIVOTAL PHASE (3)PHASE 1Pre-IND
CQADevelopment
(QbDProcess)
SpecsLifeCycle
Mgmt
CMCandTech
TransferProcess
Analytical
Manufacturing
Strategic or
Tactical Changes
Method
qualification
Dose
change
Delivery
Device
PQ
lots
Setting of Initial
Specifications
Mfg
Transfer
Method
validation
Method
transfer
Formulation
Change
Process Verification
Method Maintenance
Global
Supply
Method
Change
Accelerated Development
From: Krause, S., CaSSS CMC Strategy Forum, 27Jan14, Washington, DC.
7. Accelerated CQA Development, CMC Changes, and Specifications
7
From: Krause, S., CaSSS CMC Strategy Forum, 27Jan14, Washington, DC.
FTIH POC BLA
Tox Studies
Phase 1
Phase 3
Clinical Resupply
Mfg/Formulation Change(s)
Specifications
Revision(s)
Commercial
Specifications
QTPP
Final CQAs &
Control Strategy Approval
Potential CQAs
Product & Process
Design
Life-Cycle
Management
POST-APPROVAL
CHANGES
PIVOTAL PHASE (3)PHASE 1Pre-IND
CQADevelopment
(QbDProcess)
SpecsLifeCycle
Mgmt
CMCandTech
TransferProcess
Analytical
Manufacturing
Strategic or
Tactical Changes
Method
qualification
Dose
change
Delivery
Device
PQ
lots
Setting of Initial
Specifications
Mfg
Transfer
Method
validation
Method
transfer
Formulation
Change
Process Verification
Method Maintenance
Global
Supply
Method
Change
Accelerated Development
Comp
Lots
PQ lots
Comp
Lots=
8. Accelerated CQA Development, CMC Changes, and Specifications
8
From: Krause, S., CaSSS CMC Strategy Forum, 27Jan14, Washington, DC.
FTIH POC BLA
Tox Studies
Phase 1
Phase 3
Clinical Resupply
Mfg/Formulation Change(s)
Specifications
Revision(s)
Commercial
Specifications
QTPP
Final CQAs &
Control Strategy Approval
Potential CQAs
Product & Process
Design
Life-Cycle
Management
POST-APPROVAL
CHANGES
PIVOTAL PHASE (3)PHASE 1Pre-IND
CQADevelopment
(QbDProcess)
SpecsLifeCycle
Mgmt
CMCandTech
TransferProcess
Analytical
Manufacturing
Strategic or
Tactical Changes
Method
qualification
Dose
change
Delivery
Device
PQ
lots
Setting of Initial
Specifications
Mfg
Transfer
Method
validation
Method
transfer
Formulation
Change
Process Verification
Method Maintenance
Global
Supply
Method
Change
Accelerated Development
Comp
Lots
PQ lots
Comp
Lots=
RP vs. BP (Tier 1-3)
Covered:
RP vs. BP
results
9. 9
Analytical Method
Selection
Pharmaceutical Development Supporting Studies:
Process characterization
Product characterization
Process validation
Routine Testing (registered methods):
Raw materials
In-process
Release
Stability
Intended Use
(defined)
AMD
Studies
AMD
Studies
AMQ Report
AMQ Report
Intended Use
(re-defined)
AMV Report
Identity
Safety
Purity
Quality
Potency
Quality Target Product Profile (QTPP)
Critical Quality Attributes (CQA)
Critical Process Parameters (CPP)
Krause et al., PDA TR 57, 2012
Analytical Method Development
10. Analytical Method Development
10
Analytical Method
Selection
Pharmaceutical Development Supporting Studies:
Process characterization
Product characterization
Process validation
Tier 3 Analytical Similarity
Routine Testing (registered methods):
Tier 1 Analytical Similarity
Tier 2 Analytical Similarity
Release/Stability
Raw materials
Intended Use
(defined)
AMD
Studies
AMD
Studies
AMQ Report
AMQ Report
Intended Use
(re-defined)
AMV Report
Identity
Safety
Purity
Quality
Potency
Quality Target Product Profile (QTPP)
Tier 1-3 Analytical Similarity
Critical Process Parameters (CPP)
11. Typical Analytical Method Lifecycle
11
DS/DP
Specification
Test Methods
for New Method
Robustness
Studies
Execution
QC
Dev.
AMV
Studies
(QC-Comm.)
Start PV
Stage 2
(PQ Lots)
Completed
In progress
Not started
AMV completed
Maintenance
(QC-Comm.)
Robustness
Studies
Master Plan
AMT
Studies
(QC-Dev. &
QC-Comm.)
SOP-specific
Min/Max Method
Conditions
(for PB Design)
Commercial
Specifications
Not Parallel Step
Process Color
Legend:
Method
Qualified
(SOP Lock)
12. Typical Analytical Method Lifecycle (=> BLA Submission)
12
DS/DP
Specification
Test Methods
for New Method
Robustness
Studies
Execution
QC
Dev.
AMV
Studies
(QC-Comm.)
Start PV
Stage 2
(PQ Lots)
Completed
In progress
Not started
AMV completed
Maintenance
(QC-Comm.)
Robustness
Studies
Master Plan
AMT
Studies
(QC-Dev. &
QC-Comm.)
SOP-specific
Min/Max Method
Conditions
(for PB Design)
Commercial
Specifications
Not Parallel Step
Process Color
Legend:
Method
Qualified
(SOP Lock)
13. Analytical Method Lifecycle
APT Opportunities following AMV Study Completion and BLA Approval
13
DS/DP
Specification
Test Methods for
Same SOP and
New Product
Robustness
Studies
Execution
QC
Dev.
AMV
Studies
(QC-Comm.)
PQ Lots
Mfg
Completed
In progress
Not started
AMV completed
Maintenance
AMM
(QC-Comm.)
Robustness
Studies
Master Plan
AMT
Studies
(QC-Dev. &
QC-Comm.)
SOP-specific
Min/Max Method
Conditions
(for PB Design)
Commercial
Specifications
Not Parallel Step
APT Method
AMV and AMM
(QC)
Analytical Platform
Technology
APT
Method
Robustness
and AMT
Process Color
Legend:
Method
Qualified
(SOP Lock)
APT
Method
AMQ
Krause, S., CDER/OBP Training Presentation, White Oak, MD, 28Sep15.
14. Analytical Method Lifecycle for Accelerated Programs
Additional APT Opportunities (also for Tier 3)
14
Qualification of
Test Methods
Process and/or
Product
Characterization
Representative
Samples
Available (Dev.)
Execution Reqs:
(1. IOQ Instrument)
(2. Analyst Training)
3. Final SOP version
QC Dev. or
QC Comm.
Confirm
Method
Suitability
Start PV
Stage 2
(PQ Lots)
Qualify (as relevant):
A. Accuracy/Matching
B. Precision/Reliability
C. Specificity
D. DL or QL
Qualification
Report(s)
Method
Qualification
Master Plan
Final PV Process Ranges and/or
Analytical Control Strategy
APT (Reduced)
Qualification
Opportunity
Completed
In progress
Not started
AMV completed
Not Parallel Step
Analytical Platform
Technology
Process Color
Legend:
Krause, S., CDER/OBP Training Presentation, White Oak, MD, 28Sep15.
15. Typical Specifications and Test Methods for Drug Substance for Process
Qualification (PV Stage 2)
15
Test /
Specification
Analytical Method
Status
Typical Specifications for PQ
Reported Results
Example
Appearance Compendia
Clarity: NMT EP RS III
Color: NMT Y5
Particles: Free from or practically free from visible
particles
Clarity: EP RS I
Color: Y7
Particles: Free from visible particles
Total protein APT nominal value ± 10.0% 52.0 mg/mL
cIEF APT
Peak pattern consistent with Reference Standard
Monomer: NLT 65%
Total acidic peaks: NMT 30%
Peak pattern consistent with
Reference Standard;
Monomer: 72%
Total acidic peaks: 14%
Target binding
bioassay
Qualified / Validated N/A 98%
MOA-simulated
bioassay
Validated
90-120% (symmetrical) of Reference Standard binding
90-125% (geometrical) of Reference Standard binding
98%
Reducing gel
electrophoresis
APT
Area percent purity of heavy + light chains: NLT 98.5%
Total area percent of impurities: NMT 1.5%
Area percent purity of heavy + light
chains: 99.2%
Total area percent of impurities:
0.8%
Non-reducing gel
electrophoresis
APT
Major product peak: NLT 98.5%
Total area percent of impurities: NMT 1.5%
Major product peak: 99.2%;
Total area percent of impurities:
0.8%
HPSEC APT
Major product peak: NLT 98.3%;
Aggregates: NMT 1.7%
Fragments: NMT 1.7%
Major product peak: 99.2.0%
Aggregates: 0.5%
Fragments: 0.3%
Host cell DNA APT LT 20 pg DNA/mg protein 2 pg DNA/mg protein
CHO host cell
protein
APT NMT 20 ng/mg protein 2 ng/mg protein
Protein A APT NMT 10 ng/mg protein 1 ng/mg protein
Bioburden Compendia NMT 10 CFU per 100 mL 0 CFU per 100 mL
Endotoxin (LAL) Compendia NMT 0.20 EU/mg protein 0.01 EU/mg protein
16. Brief Summary from OBP/CDER Presentation – White Oak (28Sep15)
16
• APT status can be extended to additional products if the test method remains
essentially unchanged through proper use of change control.
• Similar to compendial method verification, “approved” test methods can be verified
(versus validated) for use of additional products.
• Reduced method qualification and transfer studies could then be executed, provided
the same sending and receiving units are used.
• For additional product license applications, the sponsor should resubmit the initial
method validation study report together with product-specific verification results
(and reports).
• For separation tests, an actual manufacturing batch should be used for test system
control as test result drift and/or variation could be observed over time; degraded
batch samples closer to the out-of-specification (OOS) level could provide better
system suitability control, and they may also provide more confidence in the test
results when results are close to the OOS level.
Krause, S., PDA Letter, May 2016
Krause, S., PDA J Pharm Sci Tech, (TBD) 2016
17. 17
Outline
1) Review of strategic opportunities to reduce analytical method lifecycle
steps for accelerated programs
- Analytical platform technology (APT) methods for Tier 2 category
- Product and process characterization APT methods for Tier 3
category
- APT conditions
2) Review of need for increased analytical method performance
- Difference(s) and challenges for Biosimilars
- AMV acceptance criteria for Tier 1 and 2
17
21. Biosimilar Analytical Method and Specification Lifecycle(s)
Use of APT Method Concept
21
AMV
Studies
Analytical
Similarity
Studies
Maintenance
(continuous
AMV)
AMT
Studies
(if needed)
PQ (PV Stage 2)
Studies
Method
Qualified
Pivotal/Phase 3
Specifications
Initial
Specifications
Tier 1 Equivalence
Tier 2 Quality Range
Tier 3 Methods
Proposed Commercial
Specifications
APT
Verification
Continuous method
verification (stage 3)
already ongoing
APT
Verification
Not needed
if SU and
RU
unchanged
22. Typical Risk-Based AMV Acceptance Criteria
22
Specifications
Consider Type
of
Specifications
Acceptance
Criteria
Existing
Knowledge
One-Sided
Specifications
(NMT, NLT, LT)
Two-Sided
Specifications
(Range)
Regulatory
Requirements
Historical
Method
Performance
Historical Data
from this
Product and
Process
Knowledge
from Similar
Product and
Process
Krause et al., PDA TR 57, 2012.
23. Tier 1 and 2 Risk-Based AMV Acceptance Criteria
23
Available RP
Lot Variation
and Mean
Consider Type
of
Specifications
Acceptance
Criteria
Existing
Knowledge
One-Sided
Specifications
(NMT, NLT, LT)
Two-Sided
Specifications
(Range)
Regulatory
Requirements
Historical
Method
Performance
Historical Data
from this
Product and
Process
Knowledge
from Similar
Product and
Process
24. Tier 1 and 2 Risk-Based AMV Acceptance Criteria
24
Available RP
Lot Variation
and Mean
Consider Type
of
Specifications
Acceptance
Criteria
Existing
Knowledge
One-Sided
Specifications
(NMT, NLT, LT)
Two-Sided
Specifications
(Range)
Regulatory
Requirements
Historical
Method
Performance
Historical Data
from this
Product and
Process
Knowledge
from Similar
Product and
Process
Tighter AMV
acceptance
criteria needed
for Tier 1 or
Tier 2
25. Tier 1 and 2 Risk-Based AMV Acceptance Criteria
25
Available RP
Lot Variation
and Mean
Consider Type
of
Specifications
Acceptance
Criteria
Existing
Knowledge
One-Sided
Specifications
(NMT, NLT, LT)
Two-Sided
Specifications
(Range)
Regulatory
Requirements
Historical
Method
Performance
Historical Data
from this
Product and
Process
Knowledge
from Similar
Product and
Process
Tighter AMV
acceptance
criteria needed
for Tier 1 or
Tier 2
Tier 1
equivalence
or Tier 2
quality
range
26. Tier 1 and 2 Risk-Based AMV Acceptance Criteria
26
Available RP
Lot Variation
and Mean
Consider Type
of
Specifications
Acceptance
Criteria
Existing
Knowledge
One-Sided
Specifications
(NMT, NLT, LT)
Two-Sided
Specifications
(Range)
Regulatory
Requirements
Historical
Method
Performance
Historical Data
from this
Product and
Process
Knowledge
from Similar
Product and
Process
Tighter AMV
acceptance
criteria needed
for Tier 1 or
Tier 2
Tier 1
equivalence
or Tier 2
quality
range
This part
essentially
goes away
for AMV
stage 2
27. Tier 1 and 2 Risk-Based AMV Acceptance Criteria
27
Available RP
Lot Variation
and Mean
Consider Type
of
Specifications
Acceptance
Criteria
Existing
Knowledge
One-Sided
Specifications
(NMT, NLT, LT)
Two-Sided
Specifications
(Range)
Regulatory
Requirements
Historical
Method
Performance
Historical Data
from this
Product and
Process
Knowledge
from Similar
Product and
Process
Tighter AMV
acceptance
criteria needed
for Tier 1 or
Tier 2
Tier 1
equivalence
or Tier 2
quality
range
This part
essentially
goes away
for AMV
stage 2
Critical review for APT methods
performance (AMV stages 2 and 3)
28. Tier 1 and 2 Risk-Based AMV Acceptance Criteria
28
Available RP
Lot Variation
and Mean
Consider Type
of
Specifications
Acceptance
Criteria
Existing
Knowledge
One-Sided
Specifications
(NMT, NLT, LT)
Two-Sided
Specifications
(Range)
Regulatory
Requirements
Historical
Method
Performance
Historical Data
from this
Product and
Process
Knowledge
from Similar
Product and
Process
Tighter AMV
acceptance
criteria needed
for Tier 1 or
Tier 2
Tier 1
equivalence
or Tier 2
quality
range
This part
essentially
goes away
for AMV
stage 2
Critical review for APT methods
performance (AMV stages 2 and 3)
- which AMV stage 2 parts not repeated ?
29. 29
EquivalenceLimit
- 1.5x RP SD 0 + 1.5x RP SD
EquivalenceLimit
No difference
Tier 1 Equivalence Testing
RP SD = Reference product variation (standard deviation) as tested
(Equivalence Demonstrated)
RPSpecifications
RPSpecifications
90% 2-sided CI for ∆ (RP-BP)
Chow, SC, “On Assessment of Analytical Similarity in Biosimilar Studies.” Drug Des, 2014.
30. Tier 1 Equivalence Testing
• RP lots (n) tested for equivalence ≈ BP lots (n)
- Example n=10 RP lots and n=10 BP lots
• Use available n=20 RP lots
- Randomly select n=10 RP lots (1st set) to set equivalence limits (+/- 1.5 RP SDs)
- Use remaining n=10 RP lots (2nd set) to test for equivalence vs. n=10 BP lots
- Mean/SD of 1st set of RP lots (n=10) ≈ 2nd set
• RP lot SDobserved = SQR [(RP lot SDtrue)2 + (RP SDanalytical)2]
• RP SDanalytical ≈ BP SDanalytical
• Tier 1 potency test considerations:
- Is RP and BP dosing done with potency units or total protein ?
- Is there a “gold” reference standard ?
Assumptions and Conditions => AMV Acceptance Criteria
30
Chow, SC, “On Assessment of Analytical Similarity in Biosimilar Studies.” Drug Des, 2014.
31. Simulated Tier 1 Example
31
EquivalenceLimit
- 1.5x RP SD 0 + 1.5x RP SD
EquivalenceLimit
No difference
RP 1.0 SDs (n=10)
32. Simulated Tier 1 Example
RP – BP = 0.5 SD
32
EquivalenceLimit
- 1.5x RP SD
0
+ 1.5x RP SD
EquivalenceLimit
RP 1.0 SDs (n=10)
BP 1.0 SDs (n=10)
90% CI RP-BP
33. Simulated Tier 1 Example calculated
RP – BP = 0.5 SD
33
EquivalenceLimit
- 1.5x RP SD
0
+ 1.5x RP SD
EquivalenceLimit
RP 1.0 SDs (n=10)
BP 1.0 SDs (n=10)
90% CI RP-BP
RP lots 1-10 = RP lots 11-20
= [(1.0)2 + (1.0)2] / 2 = 1.0
= = sqr [(2(1.0) / 10] = 0.447
Lower 90% CI Limit = 0.50 - (1.734)(0.447) = -0.275
Upper 90% CI Limit = 0.50 + (1.734)(0.447) = +1.275
34. Simulated Tier 1 Example calculated
RP – BP = 0.5 SD
34
EquivalenceLimit
- 1.5x RP SD
0
+ 1.5x RP SD
EquivalenceLimit
RP 1.0 SDs (n=10)
BP 1.0 SDs (n=10)
90% CI RP-BP
RP lots 1-10 = RP lots 11-20
= [(1.0)2 + (1.0)2] / 2 = 1.0
= = sqr [(2(1.0) / 10] = 0.447
Lower 90% CI Limit = 0.50 - (1.734)(0.447) = -0.275
Upper 90% CI Limit = 0.50 + (1.734)(0.447) = +1.275
35. Simulated Tier 1 Example calculated
RP – BP = 0.5 SD
35
EquivalenceLimit
- 1.5x RP SD
0
+ 1.5x RP SD
EquivalenceLimit
RP 1.0 SDs (n=10)
BP 1.0 SDs (n=10)
90% CI RP-BP
RP lots 1-10 = RP lots 11-20
Note: No advantage of using lots 1-10 for +/- limits
= [(1.0)2 + (1.0)2] / 2 = 1.0
= = sqr [(2(1.0) / 10] = 0.447
Lower 90% CI Limit = 0.50 - (1.734)(0.447) = -0.275
Upper 90% CI Limit = 0.50 + (1.734)(0.447) = +1.275
36. Tier 1 Equivalence Testing
• RP lot SDobserved = SQR [(RP lot SDtrue)2 + (RP SDanalytical)2]
• RP SDanalytical ≈ BP SDanalytical
• Typical AMV acceptance criteria example: Use product specifications (no potency
loss over DP shelf-life):
• If (RP and) BP potency specs are 80-125% (geometrical) or 80-120% (symmetrical),
and BP CPK = 1.00: AMV Int. Precision to be NMT 1.0 SD (7%)
• Biosimilar Tier 1 AMV acceptance criteria example(s) for analytical similarity:
- AMV Int. Precision RP/BP NMT 0.7 SD (in %CV) of RP lot SDobserved (assuming CPK
≥ 1.00): Approx. NMT 5%
- RP lot SDobserved = 1.0 = SQR [(RP lot SDtrue = 0.7)2 + (RP SDanalytical = 0.7)2] = 0.49 + 0.49
= 1.0
• AMV Accuracy and/or Specificity (“matching” or “recovery”) acceptance criteria to be
for mean difference NMT +/- 0.2 SD for BP (vs. RP)
Assumptions and Conditions => AMV Acceptance Criteria
How “good” does my (potency) test performance need to be ?
36
37. Additional Points to Consider for Tier 1
N = 10 for both reference and test product
𝜟 = 𝟏. 𝟓 𝒔 𝑹
Data from Gaussian distribution with 𝜎 𝑇𝑜𝑡𝑎𝑙 =
𝜎 𝑃𝑟𝑜𝑑
2
+ 𝜎𝐴𝑛
2
, where 𝜎 𝑃𝑟𝑜𝑑
2
and 𝜎𝐴𝑛
2
are the true
variance values for the manufactured lots and for the
analytical method, respectively.
- Only about 82% expected pass rate when the
mean difference is truly equal to zero (products are
exactly biosimilar). The 1.5 multiplier (𝒔 𝑹) may be
too small ?
- For our example (0.5 SDs true difference), pass
rate is 63%. The 1.5 multiplier may be too small.
- When the analytical variation is very high
relative to product variability, the tier 1 pass
rate remains relatively high for both small and
large product mean differences. When
analytical variation is high relative to the total
standard deviation, true product differences can be
obscured. Manufacturer’s risk is greatest when
it pursues to lower analytical variation.
Courtesy of Steven Novick, MedImmune, USA37
∆ of 0 => 82% pass
∆ of 0.5 => 63% pass
38. Tier 2 Quality Range Testing (One-Sided)
Assumptions and Conditions => AMV Acceptance Criteria
How “good” does my potency test performance need to be ?
38
+ 2.0x SD of
RP Mean
QualityRangeLimit
RPMean(n=20)
RP 1.0 SDs
(n=20)
0.0%
Impurity
39. Tier 2 Quality Range Testing
• A one-sided (non-inferiority) test model is used for a one-sided
specification (NMT X.X% impurity).
• Use all available n=20 RP lots to set mean + 2.0 SD limit for quality range.
- Data transformation is used as data is not symmetrical.
- 90% of BP results (9/10) are expected to fall within 2.0 SD limit.
(2/100 (2%) are expected to be > 2.0 SD limit if RP mean/SD = BP mean/SD)
• RP lots (n) tested for equivalence > BP lots (n)
- Example n=20 RP lots and n=10 BP lots
- No side-by-side testing (to reduce analytical variation)
• RP lot SDobserved = SQR [(RP lot SDtrue)2 + (RP SDanalytical)2]
• RP SDanalytical ≈ BP SDanalytical
Assumptions and Conditions => AMV Acceptance Criteria
39
40. Tier 2 Quality Range Testing
• RP lot SDobserved = SQR [(RP lot SDtrue)2 + (RP SDanalytical)2]
• RP SDanalytical ≈ BP SDanalytical
• Typical AMV acceptance criteria example: Use product specifications (Impurity
NMT 2.0%)
• If (RP and) and BP CPK = 1.00: AMV Int. Precision to be NMT 1.0 SD (ex., 0.5%)
• Biosimilar Tier 2 AMV acceptance criteria example(s): For analytical similarity:
AMV Int. Precision RP/BP NMT 0.7 SD (in %CV) of RP lot SDobserved (likely CPK ≥
1.00), therefore, NMT 0.35%
• AMV Accuracy and/or Specificity acceptance criteria to be for mean difference NMT
+/- 0.2 SD (ex., 0.1%) for RP vs. BP.
• Note: Conceptually, the same issues arise as those noted for Tier 1.
Assumptions and Conditions => AMV Acceptance Criteria
How “good” does my analytical method performance need to be ?
40
41. Summary of Biosimilar AMV
41
• To be fast and successful, we should use an APT concept for
Biosimilars.
• More stringent method performance expectations (ideally) exist for
AMV (Tier 1 and 2 methods) although the current Tier 1 and 2
analytical similarity acceptance criteria may not be in the interest of
patients.
• The most critical CMC success step is passing the formal analytical
similarity study (not PPQ).
Krause, S., PDA J Pharm Sci Tech, (TBD) 2016
43. 43
HPSEC AMV Study Example for an Analytical
Platform Technology (APT) Method
44. AMV Study Example(s)
Purity by HPSEC – Initial AMV Study
44
M%
A%
F%
Spiking highly degraded
product
(A% and F%)
Demonstrate:
- QLs for A% and F%
- Peak ID
- Peak separation
- Accuracy (expected peak
recoveries)
- Other AMV parameters:
Linearity, range, precision
levels
(Robustness completed before
AMV)
45. AMV Study Tier 2 Example
Post-BLA APT AMV Study
45
M%
A%
F%
For APT method:
- Continue use of validated
sample preparation
- Same assay control and system
suitability conditions
Prior to APT method verification:
- historical method performance
data from other product(s)
- product-specific data
For APT method verification:
- Run limited spiking study with
degraded product to confirm:
- QLs for A% and F%
- Peak ID
- Peak separation
- Accuracy/specificity for all
spiked levels
- Repeatability precision
- Use formal verification protocol
and justified acceptance criteria
46. 46
Typical AMV Execution Matrixfor APT Methods
(ex. HPSEC - Quantitative Limit Test)
ICH Q2(R1)
Validation
Characteristic
Analyst
Number
Day
Number
Instrument
Number
Validation Design
(Spiked Analyte Concentration)
Accuracy 1 1 1 Spike A%/F% (to final %):
0.5, 1.0, 2.0, 4.0% (run each 3x)
Repeatability 1 1 1 From accuracy
Specificity 1 1 1 Formulation matrix interference
tested (and Inferred from
accuracy)
Linearity 1 1 1 From accuracy
Assay Range 1 1 1 From accuracy
QL 1 1 1 From accuracy
47. 47
Analytical Ultracentrifugation (AUC) AMQ Study Example
for an Analytical Platform Technology (APT) Method
(Test method is used as part of the Tier 3 analytical similarity testing
program)
48. Sedimentation Velocity (SV): a method orthogonal to SEC for
detecting protein species with different molecular weights
In SV, the time dependent morphology
changes of the protein/buffer boundary of a
sample subjected to centrifugal force are
determined by the contents of size variants.
48
49. AMQ Study Example(s)
SV Method – Initial AMQ Study
49
M%
A%(Dimer,etc.)
F%
Spiking highly degraded
product
(A% and F%)
Confirm (for quant. limit test):
- QLs for A% and F%
- Peak ID
- Peak separation
- Accuracy (expected peak
recoveries)
- Other AMQ parameters:
Linearity, range, specificity,
precision
50. AMQ Study Example(s)
Post-BLA APT AMQ Study
50
M%
A%(Dimer,etc.)
F%
For APT method:
- Continue use of qualified
sample preparation
- Same assay control and system
suitability conditions
Prior to APT method verification:
- historical method performance
data from qualified method
For APT method verification:
- Run limited spiking study with
degraded product to confirm:
- QLs for A% and F%
- Peak ID
- Peak separation
- Accuracy/specificity for all
spiked levels
- Repeatability precision
- Use product-specific verification
master plan and acceptance
criteria