On this page
What should an IVIVC report demonstrate?
An in-vitro/in-vivo correlation report should explain the intended relationship, source formulations and data, model development and evaluation, and the limits of prediction. Separate fitting the relationship from evaluating its usefulness. State the proposed application without implying that a correlation plot automatically supports a waiver or specification.
Before you begin
Reporting an in-vitro/in-vivo relationship intended to inform biopharmaceutic decisions. Acceptance of the model or a proposed waiver remains product- and application-specific.
What you will prepare: An auditable IVIVC report with identified source formulations, model development, evaluation and limits of prediction.
Separate the model question from release testing
State the intended use of the relationship: explaining formulation behavior, predicting exposure, supporting a change or informing a proposed dissolution specification. M4E distinguishes dissolution work providing BA/correlation information in 5.3.1.3 from routine quality-control or batch-release dissolution in Module 3. One dataset may inform both tasks, but the reports make different arguments.
Obtain the dissolution method and its development record, formulation/batch descriptions, relevant human BA reports, individual or aggregate source data as appropriate, model plan and evaluation outputs. If the dosage form or proposed regulatory use is unknown, resolve those facts before selecting the scientific framework. “There is a correlation plot” is not enough to establish applicability.
Build a formulation-to-prediction evidence map
As a practical authoring aid, make a row for each formulation used in development or evaluation. Identify the dissolution profile, clinical study, release characteristics, batch, analysis version and role in the model. Explain any normalization, time transformation or handling of incomplete absorption and why it is appropriate for the question.
Describe the model in sufficient detail for a qualified reviewer to reconstruct the calculations. Separate datasets used to fit the relationship from those used to examine its predictive performance. Report the actual prediction checks and their limitations; do not select an evaluation method after seeing which one yields the most favorable description.
Write the boundaries of the inference
Present observed and predicted behavior with uncertainty and explain where performance changes across formulations or release conditions. A useful conclusion identifies the range of products and conditions represented by the data, the proposed use and any extrapolation beyond that range. Do not describe a fitted curve as a universally validated substitute for clinical evidence.
Fictional exercise: a model was developed using formulations with similar release profiles and then used to support a much slower proposed product. The author should flag the untested range, show the available prediction evidence and request scientific assessment of the extension. Changing the axis scale until the profiles look close is not a validation step.
Coordinate the clinical and quality decisions
Link the final model report to the BA reports and to the quality section that uses its conclusions. Preserve the same formulation and batch identities in each document. If a proposed specification or waiver relies on the model, have the responsible owners document the separate regulatory justification and any agency advice; do not imply that saving the report grants the proposal.
Before review closes, trace one prediction to its input profile and model version, then compare its intended use with the conditions actually evaluated. A missing input is a report gap to resolve, not a value to backfill from a neighboring formulation.
Give every formulation a declared role in the model
Use a formulation ledger to keep model development, evaluation and proposed use distinct.
| Formulation role | Records to connect | Boundary to explain |
|---|---|---|
| Development | Batch, dissolution profile and human study | Which behavior informed the fitted relationship |
| Evaluation | Data source, model version and prediction output | What evaluation was actually performed |
| Proposed use | Product or change supported by the model | Whether the proposal falls within represented conditions |
| Quality application | Proposed control and scientific rationale | Which separate decision uses the model result |
Review exercise: a formulation originally described as an evaluation dataset was used to modify the model after its results were inspected. Preserve that development history and have the scientist reassess how the predictive evidence should be described. Calling it independent validation after the fact would misrepresent its role.
Trace one prediction to the actual input profile and controlled model version. Link the underlying BA report and the pharmaceutical development section that uses the result. An unexplained mismatch in batch identity can undermine the bridge even when the curve is reproduced correctly.
Your preparation checklist
0/4 checkedUse this to track your review in this visit. Checks are not saved and do not establish regulatory compliance.
Frequently asked questions
Is routine release dissolution testing an IVIVC report?
No. M4E distinguishes routine quality-control dissolution in Module 3 from work providing bioavailability or correlation information in 5.3.1.3. A dataset can contribute to both purposes, but the reports explain different questions and need the appropriate evidence and cross-references.
Does a good model fit establish reliable prediction for another formulation?
Not alone. Describe the evaluation actually performed and the range of formulations and conditions represented. A proposed use outside that range needs a scientific assessment of extrapolation and uncertainty. A visually close fit to development data is not that assessment.
Does an IVIVC report automatically grant a biowaiver?
No. A model can contribute evidence to a proposal, but its adequacy and regulatory use depend on the product, context and applicable framework. Keep the report’s measured and modeled findings separate from the waiver justification and any actual agency decision.
Sources and revisions
Requirements, source recommendations and editorial preparation advice have different roles. Review the scope and revision of the source you use.
Guidance
FDA M4E(R2): The CTD: Efficacy ↗July 2017, Revision 1, final. Module 5, printed pp.56–64. Organization guidance, not a list of studies required for every application. Reopened September 22, 2026.
Guidance
FDA Bioavailability Studies Submitted in NDAs or INDs ↗April 2022 final. Study-design considerations and in-vitro approaches; NDA/IND scope. Reopened September 22, 2026.
Technical specification · placement only
FDA eCTD v4.0 comprehensive hierarchy ↗Version 2.2, February 2025. Section 5.3.1.3. A heading identifies placement, not mandatory applicability.