This is a ready-to-use classification matrix for every data point a CMC author cites in S.2.5, P.3.5, S.2.6, or P.2. It checks two things FDA’s August 2026 “Frequently Asked Questions, Developing Potential Cellular and Gene Therapy Products” guidance (docket FDA-2024-D-4311) states directly: whether a result is characterization evidence or release evidence, and whether a result generated at a scaled-down model is being cited for a claim that requires commercial-scale data. Run it before a draft S.2.5 or P.3.5 goes to QA and regulatory sign-off. Replace every <<FILL: ...>> placeholder with your own specifics. A filled specimen follows.
Document control header
| Field | Entry |
|---|---|
| Document title | CMC Filing Data-Source Classification Matrix for <<FILL: product, submission type and number>> |
| Document number | <<FILL>> |
| Version | <<FILL>> |
| Review date | <<FILL>> |
| Reviewer(s) | <<FILL>> |
1. Purpose
This matrix confirms that every data point cited toward a CQA identification, a comparability argument, or a PPQ conclusion in the drug substance or drug product sections of Module 3 is classified correctly by testing type (characterization or release) and by manufacturing scale (scaled-down model or commercial scale), and that each is cited only where that classification supports the claim being made.
2. Scope
Applies to <<FILL: submission type, e.g. original BLA, supplement>> for <<FILL: product>>, covering every data citation in 3.2.S.2.5, 3.2.P.3.5, 3.2.S.2.6, and 3.2.P.2, and their Quality Overall Summary counterparts.
3. Classification reference
| Dimension | Category | What it means | Where it may be cited |
|---|---|---|---|
| Testing type | Characterization | Generates information about the product; no predefined acceptance criteria; not on the certificate of analysis | S.2.6, P.2, comparability and CQA-identification narrative |
| Testing type | Release | Predefined numerical acceptance criteria; must comply with current good manufacturing practice; appears on the certificate of analysis | S.4.1 (specification), S.2.5 and P.3.5 batch disposition claims |
| Manufacturing scale | Scaled-down model | Evaluating process variability, setting process parameters, supportive studies such as viral or impurity clearance | S.2.6, P.2, design and characterization narrative only |
| Manufacturing scale | Commercial scale | The intended commercial process and equipment train | S.2.5 and P.3.5 PPQ conclusions; the master PPQ summary table |
4. Data-source matrix
| # | Data point cited | Section citing it | Testing type (characterization / release) | Manufacturing scale (scaled-down / commercial) | Claim it supports | Classification correct for that claim? |
|---|---|---|---|---|---|---|
| 1 | <<FILL: e.g. viral clearance spike study result>> | <<FILL: e.g. S.2.6>> | <<FILL>> | <<FILL>> | <<FILL: e.g. viral clearance capability>> | <<FILL: Yes/No>> |
| 2 | <<FILL: e.g. batch 1 potency bioassay result>> | <<FILL: e.g. S.2.5 master summary table>> | <<FILL>> | <<FILL>> | <<FILL: e.g. PPQ acceptance criterion met>> | <<FILL>> |
| 3 | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> |
Add one row per distinct data point cited toward a validation, comparability, or CQA-identification claim; a typical S.2.5 or P.3.5 section yields ten to twenty rows once the master PPQ summary table and the supporting characterization narrative are both walked line by line.
5. Misclassification disposition
| # | Row # from section 4 | Misclassification found | Correction required | Owner | Status |
|---|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL: e.g. scaled-down viral clearance data cited toward the PPQ conclusion rather than confined to S.2.6>> | <<FILL: e.g. move citation to S.2.6; cite only commercial-scale batch data in the S.2.5 PPQ conclusion>> | <<FILL>> | <<FILL: Open/Closed>> |
6. Disposition
| Field | Entry |
|---|---|
| Total data points checked | <<FILL>> |
| Misclassifications found (blocks sign-off until resolved) | <<FILL>> |
| Recommendation | Approve as correctly classified / Hold pending resolution of item(s) <<FILL>> |
7. References
FDA guidance, “Frequently Asked Questions, Developing Potential Cellular and Gene Therapy Products” (August 2026, docket FDA-2024-D-4311). See the paired article, writing CMC validation summaries for regulatory filings, for the reasoning behind this classification discipline, and the CDMO technical assessment for the same scaled-down-model restriction applied when qualifying a manufacturing partner.
Confirm the current version of the referenced guidance before issue.
8. Revision history
| Version | Date | Author | Summary of change |
|---|---|---|---|
<<FILL: 1.0>> | <<FILL>> | <<FILL>> | Initial issue. |
9. Approvals
| Role | Name | Signature | Date |
|---|---|---|---|
| Reviewer (CMC or regulatory writer) | <<FILL>> | ||
| Approver (Regulatory Affairs Lead) | <<FILL>> |
Filled specimen
Extract for an example cell therapy BLA’s S.2.5 and S.2.6 content. Company and product are illustrative.
| # | Data point cited | Section citing it | Testing type | Manufacturing scale | Claim it supports | Classification correct? |
|---|---|---|---|---|---|---|
| 1 | Bench-scale bioreactor viral clearance spike study, log reduction value | S.2.6, viral safety narrative | Characterization | Scaled-down | Viral clearance capability of the purification step | Yes, correctly confined to S.2.6 and not cited in the S.2.5 PPQ conclusion |
| 2 | Batch 2 potency bioassay result, 105 percent relative activity | S.2.5 master PPQ summary table | Release | Commercial | PPQ acceptance criterion met for potency | Yes |
| 3 | Small-scale resin lifetime study, cycles to failure | Originally cited in S.2.5 as supporting the PPQ conclusion on process consistency | Characterization | Scaled-down | PPQ conclusion on process consistency | No, scaled-down characterization data cited toward a PPQ claim; commercial-scale resin lifetime data does not yet exist for this campaign |
Disposition on row 3: misclassification found. The scaled-down resin lifetime study is scientifically sound evidence for the process design narrative in S.2.6, but FDA’s August 2026 FAQ guidance is specific that PPQ conclusions need commercial-scale data. Correction: the citation moved to S.2.6’s process design discussion, and the S.2.5 PPQ conclusion was rewritten to state resin lifetime is being confirmed through the Stage 3 continued process verification program rather than asserted as already demonstrated at commercial scale.
Common findings this matrix prevents
- A PPQ conclusion supported in part by scaled-down or bench-scale data, discovered only when a reviewer asks where the cited study was actually run.
- A characterization result presented in a way that reads as though it were tested against a release specification, with no acceptance criterion stated because none exists for that result.
- A comparability argument built on a mix of characterization and release data with no statement of which is which, so the reviewer cannot judge the strength of the claim.
- A CDMO-supplied data package that was never checked for scale before the sponsor’s own narrative cited it toward a commercial-scale claim.
How to adapt this matrix
- Run this matrix after S.2.5, P.3.5, S.2.6, and P.2 drafts exist, not before, since classifying citations against a still-changing narrative wastes the exercise.
- Walk the master PPQ summary table and the characterization narrative line by line; do not rely on memory of which study supported which claim.
- Route every misclassification found in section 5 back to the drafting author for correction, not a rewording that keeps the same citation in the same place.
- Re-run this matrix if either narrative changes after this review, since a citation moved between sections can silently break its own supporting logic.