Home/Services/EU MDR Clinical Evaluation/Clinical Evaluation Reports for SaMD and AI Medical Devices

Clinical Evaluation Reports for SaMD and AI Medical Devices

Clinical Evaluation Reports for SaMD and AI Medical Devices

SaMD clinical evaluation report writing for software and AI-enabled medical devices: intended purpose, clinical evidence, performance and version traceability.

The IMTS perspective

Evidence that stands up to scrutiny.

IMTS writes clinical evaluation documentation for software as a medical device, including AI-enabled devices. We connect the intended medical purpose, evaluated software version, clinical workflow and evidence supporting the claimed output.

A high model-performance statistic does not explain whether a software output supports its intended clinical use. The evaluation needs a traceable relationship between the claim and the population, setting, endpoints and limitations of the evidence.

Scope of support

What we can deliver.

How we work

A focused path from source evidence to final document.

01

Define the medical claim

Confirm the software version, inputs, output, intended users, population and clinical decision supported.

02

Map the evidence

Organize available studies and technical records by the evidence question they answer and identify unsupported steps.

03

Document the clinical argument

Explain applicability, uncertainty and limitations; reconcile the CER with labeling and planned follow-up.

IMTS expertise

Authorship and review.

Clinical Evaluation Reports for SaMD and AI Medical Devices

The three evidence questions for medical device software

MDCG 2020-1 distinguishes valid clinical association, technical performance and clinical performance. In practical terms, the evaluation asks whether the output has the claimed medical meaning, whether the software produces that output reliably, and whether its use achieves the intended clinical purpose. Evidence supporting one question should not be silently used to answer the others.

IMTS uses this separation to organize the document and evidence inventory. The evaluation remains device-specific. The software configuration, intended use and clinical pathway determine which records are relevant; a generic AI checklist cannot establish a sufficient clinical evidence package.

Clinical Evaluation Reports for SaMD and AI Medical Devices

AI evidence: connect reported results with the actual claim

Provide the study design, population, data provenance, validation approach, reference standard and performance results with their limitations. Identify which records apply to the released software version. If results come from a different configuration, the report needs a reasoned assessment of what can be transferred and what remains uncertain.

For an illustrative detection tool, improved lesion detection is a different claim from improved patient survival. A CER should not bridge that gap through wording alone. Review whether the observed endpoint, users and care setting support the precise benefit claimed in labeling. This example is methodological, not an IMTS client outcome.

Clinical Evaluation Reports for SaMD and AI Medical Devices

Updates and the source package we need

Changes to the software, training approach, intended population or workflow can affect the clinical argument. An update should identify the changed element and reassess the affected evidence rather than treating every version as clinically identical. The manufacturer retains responsibility for controlled change decisions and regulatory classification.

Send the intended-purpose statement, current labeling, version history, performance and clinical study reports, risk inputs and postmarket findings. Our CEP writing service can define the evaluation approach before drafting. For a new evidence period or released version, see CER update services.

Primary regulatory sources

Check the applicable consolidated legislation and current guidance for the device. Guidance supports interpretation; it does not replace the MDR or a device-specific assessment.

Connected services and working guides

For the complete evaluation package, see IMTS Clinical Evaluation Report writing services.

Discuss a SaMD clinical evaluation report with IMTS, or review the complete CER writing service.

Direct answers

Questions medical device teams ask.

Do you support AI-enabled SaMD?

Yes. IMTS supports clinical evaluation writing for AI-enabled medical device software, with the scope tied to the intended medical purpose, evaluated version and available evidence.

Can technical accuracy alone establish clinical performance?

Not automatically. The evaluation needs to assess whether the evidence supports the intended clinical purpose. Technical results and clinical evidence answer related but distinct questions.

Does this service include IVDR performance evaluation?

This page describes MDR clinical evaluation for medical device software. IVD software follows the applicable IVDR performance evaluation framework and requires separate scoping; it should not be assumed to be covered by an MDR CER.

Start a conversation

Discuss this clinical evaluation workstream.

Share the current document set, evidence period and project question so we can agree a focused scope.

Book a Scoping Call