M1: Diagnostic Function

Major Non-Conformity

Code: 2650005-202501-M1

Status: Open: awaiting manufacturer response

Requirements and references

• GSPR 1, 15.1, 17.2
• Annex II, 6.1(a), (b); 6.2(f)
• EN 62304
• EN 82304-1

Documents reviewed by BSI

• ifu.pdf
• master.csv
• C515_T355/case-001/response.json
• C537_T377/case-001/response.json

Regulatory context

Internal working document

This section is for internal planning only. It will not be included in the final response to BSI.

BSI raised four questions under Major NC 2650005-202501-M1. The regulatory requirements cited are GSPR 1, 15.1, 17.2; Annex II 6.1(a), (b), 6.2(f); EN 62304; and EN 82304-1.

Cited requirements — detailed breakdown

Each response must explicitly tie back to the specific requirement it satisfies. The cited requirements are:

Requirement	What it demands	Relevant questions
GSPR 1	Devices shall achieve intended performance, be safe and effective, with acceptable benefit-risk ratio	Q1 (performance claims must be understandable)
GSPR 15.1	Devices providing diagnostic information must indicate accuracy limits, precision, and stability; limits of accuracy shall be indicated by the manufacturer	Q1 (IFU must clearly communicate accuracy metrics and their applicability)
GSPR 17.2	Software shall be developed per state of the art: lifecycle, risk management, V&V	Q2, Q3, Q4 (V&V evidence must be complete and correct)
GSPR 23.4(c)	IFU must state residual risks and undesirable side-effects	Q1 (performance limitations must be communicated)
GSPR 23.4(e)	IFU must state warnings and precautions	Q1 (accuracy limitations per condition/context)
GSPR 23.4(h)	IFU must provide specifications for appropriate use, including degree of accuracy claimed	Q1 (core of BSI's concern — accuracy claims need context)
MDR Annex II 6.1(a)	Technical documentation must include device description and intended purpose, including performance characteristics	Q1 (performance characteristics must be documented in the technical file, not just the IFU)
MDR Annex II 6.1(b)	Technical documentation must include design and manufacturing information sufficient to understand design stages and V&V	Q2, Q3, Q4 (test environment and evidence must be traceable)
MDR Annex II 6.2(f)	Technical documentation must include verification and validation results and evidence	Q2 (test environment), Q3 (wrong evidence path), Q4 (evidence mismatch)
EN 62304 §5.5	Software integration and integration testing — requires documented test plan including environment	Q2 (test environment), Q3, Q4 (integration test evidence)
EN 62304 §5.6	Software system testing — requires documented system test plan and evidence	Q2 (test environment specification)
EN 82304-1 §6.2	Health software validation — evidence must include environment specification	Q2 (test environment)

GSPR 15.1 — our position on diagnostic function

Our GSPR analysis (R-TF-008-001) states that "the device does not have a diagnostic nor a measuring function but it is intended to provide information to support practitioners in decision making for diagnostic purposes." Despite this, GSPR 15.1 is marked as applicable because the device provides information that influences diagnostic decisions.

This position must be proactively clarified in our response. The device is a clinical decision-support tool (Rule 11, MDR Annex VIII), not a diagnostic device. It does not make diagnoses — it provides supplementary information (severity scores, ICD probability distributions, image analysis) that practitioners use alongside their clinical judgement. GSPR 15.1 applies because the accuracy and precision of that supplementary information must be communicated to users so they can weigh it appropriately. Our response should frame the IFU updates as fulfilling GSPR 15.1's requirement to "indicate limits of accuracy" for decision-support information, not as an admission that the device has a diagnostic function.

Response approach for a Major NC

This is a Major NC. Our responses need to be thorough and demonstrate corrective action. For each question:

1. State the regulatory requirement being addressed — cite specific GSPR/Annex II clauses
2. Acknowledge the gap — don't be defensive
3. Explain what was already in place — demonstrate maturity of the system
4. Describe corrective action taken — with specific document references
5. Provide evidence — red-lined documents and supplementary evidence PDF

---

Cross-cutting: evidence packaging CAPA

Systemic issue

Questions 3 and 4 both involve incorrect or mismatched evidence in the integration test package submitted to BSI. This suggests a systemic issue with evidence packaging/compilation, not isolated errors. BSI may raise this in Round 2 if we only fix the individual cases.

The systemic issue

Two out of 76 integration test evidence packages had errors:

• T355 (Q3): Evidence URI pointed to the wrong S3 path (data entry error in CSV)
• T377 (Q4): Evidence response.json does not match expected output (structural mismatch or compilation error)

This is a 2.6% error rate in evidence packaging. For a Major NC under Annex II 6.2(f), BSI will expect confidence that the remaining 74 test cases are correct.

Proposed corrective and preventive action

1. Audit all 76 evidence URIs against their test case definitions. For each test case, verify:

   • The Evidence URI path contains the correct model name
   • The response.json at that path has the same JSON structure as the expected result in the CSV
   • Numerical values are within the specified acceptance criteria

2. Add automated validation to the evidence collection pipeline:

   • Schema validation: check that evidence response.json has the same top-level keys as the expected result
   • URI validation: check that the evidence URI path matches the model name in the test case
   • Tolerance check: verify all numerical values are within acceptance criteria

3. Document the audit results as supplementary evidence (table showing all 76 test cases verified, with pass/fail and any corrections made).

Action items for CAPA

#	Action	Owner	Document affected	Priority
12	Audit all 76 integration test evidence URIs and content	Gerardo	Supplementary evidence PDF	High
13	Add automated evidence validation to integration pipeline	Gerardo	Integration test tooling	Medium

---

Summary of action items

#	Action	Owner	Document affected	Priority
1	Add clinical studies reference section to IFU	Taig + Jordi	ifu.pdf (eu-ifu-mdr)	High
2	Add ICD-11 codes to performanceClaims.ts data model	Taig + Jordi	performanceClaims.ts	High
3	Surface ICD-11 codes in IFU performance claims	Taig + Jordi	ClinicalBenefitsList.tsx, ifu.pdf	High
4	Add clinical benefit glossary/legend table to IFU	Taig	clinical-benefits-and-performance-claims.mdx	High
5	Add consolidated test environment section to Test Report	Gerardo	R-TF-012-035	High
6	Verify correct biofilm evidence exists at correct S3 path	Gerardo	—	High
7	Fix Evidence URI for T355 in CSV	Gerardo	ai-models-integration-tests.csv	High
8	Export correct biofilm/slough evidence from S3	Gerardo	Supplementary evidence PDF	High
9	Add evidence URI validation to integration test pipeline	Gerardo	Integration test tooling	High
10	BLOCKER: Investigate T377 response.json in S3	Gerardo	—	Critical
11	Provide corrected T377 evidence	Gerardo	Supplementary evidence PDF	High
12	Audit all 76 integration test evidence URIs and content	Gerardo	Supplementary evidence PDF	High
13	Add automated evidence validation to integration pipeline	Gerardo	Integration test tooling	Medium
14	Compile red-lined documents	Taig	BSI response package	High
15	Compile supplementary evidence PDF with bookmarks	Taig	BSI response package	High

Dependencies

Timeline

• Deadline: 2026-03-17 (Round 1 response)
• Blockers must be resolved by: 2026-03-05 (items 6, 10)
• Documentation updates: 2026-03-07 (items 1-5, 7)
• Evidence compilation: 2026-03-10 (items 8, 11, 12)
• Internal review deadline: 2026-03-14 (items 14, 15)
• Submission: 2026-03-17