Question for Jordi: Studies in the IFU bibliography that may raise concerns

Context

This is an internal note flagging two studies that appear in the IFU Clinical Validation Studies section (via the ClinicalStudiesReference component). These are not issues for the current M1 Q1 non-conformity, but they are likely to be raised in BSI's clinical review round. Flagging now so we can decide in advance whether to change anything.

Issue 1: MC_EVCDAO_2019 — Legacy Device

The study MC_EVCDAO_2019 ("Clinical validation study of a CAD system with artificial intelligence algorithms for early noninvasive in vivo cutaneous melanoma detection") was conducted with the Legit.Health Legacy Device, not Legit.Health Plus. The bibliography currently shows this as "Device version: Legit.Health Legacy Device."

Why this is a problem: A BSI clinical reviewer reading the IFU will ask why evidence from a different device version is being cited as supporting evidence for the current device. Without a bridging argument or equivalence determination visible in the IFU, this looks like unsupported extrapolation.

Options:

1. Exclude from the IFU bibliography. Filter ClinicalStudiesReference to only show studies with medicalDevice: "Legit.Health Plus". The study would still exist in the technical file and CER but would not appear in the user-facing IFU.
2. Add a bridging note. Include a brief statement in the IFU explaining that performance data from the predecessor device is included based on a demonstrated equivalence determination documented in the Clinical Evaluation Report (R-TF-015-003). This is more transparent but could invite further scrutiny if the equivalence argument is not airtight.
3. Do nothing. Leave it as is and address it reactively if BSI raises it in the clinical review.

Issue 2: AIHS4 2025 — Sample size of 2

The study AIHS4 2025 ("Evaluation of AIHS4 Performance in the M-27134-01 Clinical Trial for Hidradenitis Suppurativa") has a sample size of 2 participants. It appears in the IFU bibliography alongside studies with sample sizes of 29–204.

Why this is a problem: A sample size of 2 does not constitute a meaningful clinical validation study by any standard definition. Including it in the IFU alongside properly powered studies undermines the credibility of the evidence package. A BSI reviewer is likely to question whether this qualifies as clinical evidence and may use it to challenge the overall rigor of the clinical evaluation.

Options:

1. Exclude from the IFU bibliography. Set sampleSize: null in clinicalStudiesData.ts so the ClinicalStudiesReference component filters it out (the component already filters by sampleSize !== null). The study would remain in the technical file for completeness. Note: this would also remove AIHS4's performance claims (LL5, SDP) from the STUDY_IDS array and from the benefit 5RB listing — verify whether this is acceptable.
2. Reclassify as a feasibility/pilot observation. Keep the data but relabel it in the bibliography to make clear it is a preliminary observation, not a powered study. This is more honest but still looks weak.
3. Do nothing. Leave it as is and defend it if challenged.

Recommendation

For both issues, I'd lean towards Option 1 (exclude from the IFU bibliography) for the AIHS4 study — the risk-to-benefit ratio of including a 2-patient study in a user-facing document is unfavourable. For the Legacy Device study, I'd lean towards Option 2 (add a bridging note) if the equivalence determination in the CER is solid, or Option 1 if it is not.

Jordi, what do you think? Should we make these changes proactively before the clinical review round, or wait and address reactively?