ICH M7 in Real Life: Why Mutagenicity Assessment Is Still Misunderstood

One of the most persistent challenges in pharmaceutical regulatory submissions is the misapplication of ICH M7. Particularly around how experimental and in-silico data relate to each other. This article sets out the most common misinterpretations and errors encountered in practice.

ICH M7 is the guideline to assess and control the DNA-reactive impurities in pharmaceutical products. ICH M7 emphasises safety and quality risk management in establishing mutagenic impurity levels and limits. The guideline assesses the levels of mutagenic impurities that are expected to pose a risk of cancer. A DNA-reactive substance has the potential to damage DNA, trigger mutations, and ultimately cause cancer. This type of mutagenic carcinogen is usually detected by the bacterial reverse mutation assay, which is used to assess mutagenic potential. Another method is structure-based assessment, which is useful for predicting bacterial mutagenicity outcomes based on theoretical data.

ICH M7 uses data obtained through actual experiments and theoretical data, the latter generated through theoretical prediction via the (Q)SAR module. Two (Q)SAR prediction methodologies that complement each other should be applied: one rule-based and one statistical-based. The in-silico predictions and the experimental outcomes are not interchangeable.

Examples of misinterpretation between theoretical and practical:

The following examples illustrate recurring misinterpretations observed in regulatory submissions and industry practice. Understanding these errors is essential to ensure ICH M7 compliance and avoid regulatory challenges.

1. Misinterpretation 1: QSAR data is less important than experimental data. The (Q)SAR conclusion is based on two complementary models.  One rule-based and one statistical-based. If both models predict mutagenicity, the compound is treated as mutagenic, and experimental data from the Ames assay is not required in that specific scenario. However, this does not mean QSAR data is optional or subordinate.
2. Misinterpretation 2: Experimental (Ames assay) data always takes precedence over theoretical data. This is not always true. Experimental data takes precedence only when the study is valid and there are no conflicting concerns. Where data is inconclusive, of poor quality, or generated through an irrelevant experimental design, it is not considered valid and does not override QSAR.
3. Misinterpretation 3: A negative QSAR result always confirms non-mutagenicity, and a positive result always confirms mutagenicity. If QSAR predicts a negative result, concluding that the compound is non-mutagenic is a dangerous oversimplification. If predictions are uncertain, or if the compound falls outside the model’s applicability domain, experimental backup data from the Ames assay is required. Similarly, a positive QSAR result does not automatically confirm mutagenicity; additional experimental data via the Ames assay is required before drawing a conclusion.
4. Misinterpretation 4: QSAR assessment is optional and can be skipped. ICH M7 clearly states that QSAR is mandatory and both methodologies must be performed.
5. Misinterpretation 5: Theoretical justification alone is sufficient for Option 4 control. ICH M7 Option 4 is based on the justification provided through process controls and purge knowledge. No analytical testing of the final product is required; however, when applicants rely on simple theoretical calculations assuming a high degree of purge without supporting data the justification can be challenged by regulators. Scientific justification with supporting data is required to demonstrate that impurities remain continuously below the acceptable range when process control and purging are applied correctly.
6. Misinterpretation 6: A negative (Q)SAR result overrules a positive Ames test. When in-silico assessment using both QSAR models predicts that an impurity is non-mutagenic, but an Ames assay returns a positive result, it is a common error to dismiss the Ames result as a false positive on the basis that the theoretical model indicates safety. In practice, a positive Ames result requires full assessment and control measures; the Ames result cannot be ignored.

Common errors in Ames justification:

Common errors in Ames justification arise when organisations attempt to use experimental data or expert reasoning to overrule the QSAR outcome, or to justify the safety of an impurity without sufficient evidence. Some of the most frequently observed errors are described below.

Inadequate protocol for Ames test:

The most significant error in Ames testing is failure to follow OECD 471 guidelines strictly. In such cases, laboratories fail to use the five-strain battery, which is critical for detecting diverse mutational mechanisms. Additionally, only standard metabolic activation is used, without including specialised metabolic activation conditions or pre-incubation steps, which may be required to detect certain classes of promutagen.

Wrong interpretation or classification of Class 4 analogues:

As per ICH M7, Class 4 covers impurities that share the same structural alert as the drug substance, which itself tested negative, and on that basis the impurity can be classified as Class 4. An error occurs when organisations overlook the fact that an impurity may have a different electronic and steric environment from the parent drug substance, which can affect mutagenic potential independently of the shared alert. Organisations also err by ignoring additional, separate structural alerts when proposing Class 4 classification.

Failures in regulatory reporting:

Organisations frequently submit only conclusions from the Ames test or the QSAR report, without the underlying data or rationale.

In general, organisations fail to provide strong justification for the Ames test, including: a study protocol prepared exactly as per the OECD 471 GLP guidelines; correct interpretation of results; in-depth structural alert analysis; QSAR predictions with details of the applicability domain used; and clear classification of the impurity within ICH M7 Classes 1 to 5.

Why These Errors Matter: Regulatory Consequences

Errors in ICH M7 assessments are not merely academic. Regulatory agencies including the EMA and FDA have issued deficiency letters and clinical hold notices where mutagenic impurity assessments were found to be inadequate. Submissions that fail to provide complete QSAR reports with applicability domain details, or that dismiss positive Ames results without adequate scientific justification, are routinely challenged during review.

Role of impurity reference standards:

Impurity reference standards play a critical role in identifying, quantifying, and justifying the safety of impurities. Their critical functions include analytical testing, toxicological and hazard assessment, and quality control and regulatory compliance. In analytical testing, reference standards are used to confirm the identity of impurities, for exact quantitation, and for method validation purposes. In toxicological assessment, impurities are used to perform the Ames assay, to cross-check and confirm the predictions generated by QSAR. Reference standards can also be used in spiking studies to demonstrate that impurities are detected reliably. Impurity standards with ISO certifications are the regulatory requirements for traceability. In QC, impurity reference standards are used to monitor stability studies and batch release.

Conclusion

Getting ICH M7 right requires treating QSAR and Ames data as genuinely complementary tools. Neither is optional, and neither automatically overrides the other. The most common failures in regulatory submissions stem from shortcuts in one or both of these methods: inadequate Ames protocols, over-reliance on in-silico predictions, poorly justified Option 4 controls, or incomplete regulatory packages.

Regulators are increasingly sophisticated in their review of mutagenic impurity assessments. Submissions that treat ICH M7 as a box-ticking exercise, rather than a genuine scientific risk management framework, are likely to attract queries and delays. A thorough, integrated approach that combines robust experimental data, transparent QSAR reporting, and well-justified classification decisions is the most effective way to ensure first-cycle approval.

FAQ:

REFERENCE:

• “ICH M7 Assessment and Control of DNA Reactive (Mutagenic) Impurities in Pharmaceuticals to Limit Potential Carcinogenic Risk – Scientific Guideline | European Medicines Agency (EMA).” European Medicines Agency (EMA), 24 Sept. 2014, www.ema.europa.eu/en/ich-m7-assessment-control-dna-reactive-mutagenic-impurities-pharmaceuticals-limit-potential-carcinogenic-risk-scientific-guideline.
• ‌Amberg, Alexander, et al. “Principles and Procedures for Implementation of ICH M7 Recommended (Q)SAR Analyses.” Regulatory Toxicology and Pharmacology, vol. 77, Elsevier BV, June 2016, pp. 13–24, https://doi.org/10.1016/j.yrtph.2016.02.004.
• “(Q)SAR Assessment Framework: Guidance for the Regulatory Assessment of (Quantitative) Structure Activity Relationship Models and Predictions.” OECD Series on Testing and Assessment, OECD, Nov. 2023, https://doi.org/10.1787/d96118f6-en.