Response

Response to BSI Item 3a - Clinical Data Analysis

Following the BSI auditor's observations, a comprehensive review and update of the Clinical Evaluation Report (CER) and the supporting Clinical Investigation documentation have been completed to ensure all relevant clinical data is identified and sufficient analysis is provided.

1. Overall Analysis and Traceability The CER (R-TF-015-003) has been significantly enhanced to provide a clear, integrated analysis of how all clinical benefit, safety, and performance claims have been met:

• A new "Acceptance criteria reconciliation" section (within Section 15) has been added. This section explicitly maps every performance and safety metric from each clinical investigation against the predefined acceptance targets derived from the State of the Art (SotA). This provides direct, unambiguous traceability and a clear discussion of whether each criterion was met, along with clinical justifications for any minor deviations.
• A "Methodological justifications and statistical adequacy" section was added to transparently justify the analysis methodology. This includes the rationale for the sample sizes (e.g., explaining that the MC_EVCDAO_2019 study concluded at 105 patients because the high prevalence of malignant cases provided sufficient statistical power to validate the primary safety endpoint), the rationale for data exclusion (justifying the use of the integrated Deep Image Quality Assessment algorithm to mirror real-world use), and the coverage of both clinical and dermatoscopic imaging modalities.

2. Clinical Investigations (CIs) Documentation The regulatory and administrative details for all pivotal studies, including MC_EVCDAO_2019 and IDEI_2023, have been explicitly documented in a new "Regulatory and Administrative Details of Clinical Investigations" table in the CER. This includes:

• Competent Authority & Ethics Committee: Confirmation of AEMPS notification (observational) and specific CEIm approval reference numbers and dates.
• Public Registration: All studies have been publicly registered; their ClinicalTrials.gov (NCT) and EMA RWD Catalogue (EUPAS) numbers are now explicitly listed.
• Publication Status: The publication status in peer-reviewed journals is now tracked (e.g., BI_2024 published in JMIR Dermatology, IDEI_2023 published with DOI 10.1101/2025.03.11.25323753).
• Protocol Deviations: All protocol deviations for the pivotal studies (including the sample size adjustment in MC_EVCDAO_2019 and the expanded cohort in IDEI_2023) are now fully documented and justified within their respective Clinical Investigation Reports (R-TF-015-006). The critical deviations have also been integrated into the methodological justifications of the CER.

3. Representativeness of Patient Populations and Indications To demonstrate sufficient coverage of representative patient populations, a "Representativeness of the Study Populations (Demographics & Skin Phototypes)" table was added to the CER. This provides a comprehensive breakdown of Gender, Age distribution across life stages, and Fitzpatrick skin phototypes (I-VI) across the 800+ patients in the pivotal studies. Furthermore, the "Coverage of Indications and Conditions" analysis has been restructured around a risk-proportionate, 3-tier evidence structure: Tier 1 (malignant conditions: individual analysis with per-condition acceptance criteria), Tier 2 (rare diseases: grouped analysis with a dedicated acceptance criterion), and Tier 3 (general conditions: pooled analysis with explicit risk-based justification for pooling). This tiered approach is supported by a 7-category epidemiological framework based on the Global Burden of Skin Disease (Karimkhani et al. 2017), which demonstrates that the evidence portfolio covers 5 of 7 epidemiological categories (97% of dermatological presentations). The two categories with insufficient evidence (autoimmune diseases, 3%, and genodermatoses, 1%) are declared as acceptable gaps per MDCG 2020-6 § 6.5(e), with justification and PMCF linkage documented in R-TF-015-003, "Need for more clinical evidence."

4. Demonstration of Equivalence & Software Changes The Equivalence assessment (CER Section 16.1.4) has been thoroughly expanded. Detailed comparative tables now explicitly justify the Technical, Clinical, and Biological equivalence between the device and the legacy version.

To address the observed inconsistency between the CEP and CER regarding software changes, the documentation has been unified. The CEP and CER now explicitly list the exact improvements introduced in the current device version compared to the legacy version. These are limited strictly to software stabilization and feature consolidation required for MDR compliance:

1. Migration to a microservices architecture.
2. Implementation of the HL7 FHIR standard.
3. Database encryption and cybersecurity upgrades.
4. Enhanced user interface (UI) feedback for image quality (DIQA algorithm).

A robust justification is provided explaining that none of these changes affect the core Artificial Intelligence models, the Vision Transformer architecture, the clinical indications, or the fundamental principles of operation. Verification and validation testing has confirmed that these software stabilisation updates have not negatively impacted the clinical safety, performance, or diagnostic accuracy of the device. The new clinical data collection was primarily mandated by the regulatory shift from an MDD Class I device to an MDR Class IIb device, which requires a significantly higher level of clinical evidence.

5. Clinical Literature Search The documentation has been updated to address the clinical literature search protocol and the presentation of SotA clinical data:

• To address the confusion in Section 3.1 of the SotA, the section headers have been corrected from "Clinical data collected on..." to "State of the art data collected on...". This clarifies that these tables present the baseline performance of alternative/similar devices and standard clinical practice, rather than clinical data on the subject device itself. The formatting of the tables has also been reviewed to ensure they are not cut off.
• The CER (Section 16.1.5) has been updated with a more granular appraisal of the 13 identified publications. Of these, 8 describe purely algorithmic optimisation and benchmarking against public datasets without clinician involvement (Group A, excluded as preclinical per MDR Article 2(48)). Four are peer-reviewed validation studies comparing the device's severity assessment algorithms against expert dermatologist consensus on validated clinical scales: PASI, UAS, IHS4, and SCORAD (Group B, included as Technical Performance evidence per MDCG 2020-1 Pillar 2). The per-publication appraisal and analysis is documented in a new CER section, "Analysis of published severity validation studies." The thirteenth publication — NMSC_2025 (Medela et al., European Archives of Oto-Rhino-Laryngology 282:1585–1592, 2025, DOI: 10.1007/s00405-024-08951-z) — is a prospective diagnostic accuracy study evaluating the device's performance for non-melanoma skin cancer (BCC and cSCC) detection in a head and neck outpatient clinic (135 patients, histological confirmation for all cases; Group C, included as Clinical Performance evidence per MDCG 2020-1 Pillar 3, Rank 4 per MDCG 2020-6 Appendix III). The full appraisal and analysis is documented in a new CER section, "Appraisal of published malignancy clinical evidence."
• To address the search protocol clarity, the CER documentation now explicitly confirms: "The literature search for the device under evaluation followed the same rigorous PICO protocol and appraisal methodology as the State of the Art (SotA) search, with the only difference being the addition of specific keywords targeting publications related to the subject device and its manufacturer. This ensures that no relevant peer-reviewed clinical studies were missed."

6. Post-Market Surveillance (PMS) Data The CER Executive Summary and the PMS analysis sections have been updated to explicitly incorporate and discuss the extensive data from the market. It is now formally documented that the equivalent legacy device has generated over 4,500 clinical reports across 21 contracts since 2020, with zero complaints regarding clinical safety or performance, zero serious incidents, zero CAPAs, and zero Field Safety Corrective Actions (FSCAs), providing strong real-world confirmation of the device's long-term safety and performance profile. The use of this legacy market data as clinical evidence is grounded in MDCG 2020-6 § 6.2.2, which addresses the use of post-market data from legacy devices in the clinical evaluation of the transitioning device. The data is appraised using IMDRF MDCE WG/N56 Appendix F quality criteria.

7. Validated Methodological Quality Appraisal

A dedicated section "Validated methodological quality appraisal" has been added to the CER (R-TF-015-003), applying named, design-specific validated tools to all nine manufacturer clinical investigations in accordance with MEDDEV 2.7.1 Rev 4 Section 9 (Stage 2) and MDCG 2020-6 § 6.3.

Four study design families are present in the portfolio, each assessed with the appropriate tool:

• QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies) was applied to the four studies with an index test versus reference standard design: MC_EVCDAO_2019, IDEI_2023, AIHS4_2025, and NMSC_2025. Domain-level risk-of-bias and applicability ratings (LOW / HIGH / UNCLEAR) are documented for all four studies, with interpretive commentary on clinically relevant findings (e.g., the AUC improvement from 0.7738 to 0.9430 in IDEI_2023 following DIQA pre-processing; the Domain 1 HIGH risk of bias in NMSC_2025 reflecting the specialist referral population and the consequent 80% malignancy prevalence, which is acknowledged and does not invalidate the AUC-ROC 0.93 finding).
• MINORS (Methodological Index for Non-Randomised Studies) was applied to the three clinical utility/workflow studies: COVIDX_EVCDAO_2022 (non-comparative, 13/16), DAO_Derivación_O_2022 (comparative, 22/24), and DAO_Derivación_PH_2022 (comparative, 21/24). Scores indicate adequate to good methodological quality.
• MINORS was also applied to the three simulated-use MRMC studies: BI_2024 (22/24), SAN_2024 (22/24), and PH_2024 (23/24). These studies provide Technical Performance evidence per MDCG 2020-1 Pillar 2 and are classified as Rank 11 per MDCG 2020-6 Appendix III; they are not clinical investigations under MDR Article 2(48). The section documents the adaptation of MINORS item 2 to the MRMC study design.
• MINORS non-comparative (maximum 16) was applied to the four published severity validation studies: APASI_2025 (11/16), AUAS_2023 (11/16), AIHS4_2023 (11/16), and ASCORAD_2022 (11/16). These publications are classified as Technical Performance evidence per MDCG 2020-1 Pillar 2 and are ranked 5–6 per MDCG 2020-6 Appendix III. The 11/16 score reflects the retrospective design inherent to this study type; items 3 (prospective data collection) and 8 (prospective sample size calculation) are structurally zero for retrospective algorithm validation, and the remaining six items are all adequate.

Per-study interpretive prose — addressing study design, key findings, known limitations, and clinical significance — has been added within the "Manufacturer's clinical data" section of the CER for each of the nine studies.

8. AIHS4 Study Limitations and PMCF Mitigation: A limitation regarding the AIHS4_2025 study was formally acknowledged in the CER. While the study produced promising results for the severity assessment of Hidradenitis Suppurativa (ICC 0.727, repeated-measures design with 2 patients and 16 longitudinal assessments), the sample size was small. To address this, the PMCF Plan (R-TF-007-002) has been updated to include Activity B.5: Confirmatory validation for severity assessment of Hidradenitis Suppurativa (target sample: 100 patients) to provide the necessary confirmatory clinical evidence.

9. Justification of Sufficiency of Clinical Evidence A new dedicated section, "Justification of sufficiency of clinical evidence," has been added to the CER. This section provides an explicit justification that sufficient data in quantity and quality has been analysed. The evidence portfolio is characterised by MDCG 2020-6 Appendix III ranking rather than a uniform quality label: MC_EVCDAO_2019, COVIDX_EVCDAO_2022, and DAO_Derivación_O_2022 are classified as Rank 2 (high-quality clinical investigations with coverage gaps); IDEI_2023, DAO_Derivación_PH_2022, AIHS4_2025, and NMSC_2025 are Rank 4 (methodological limitations, data quantifiable and acceptability justifiable); BI_2024, PH_2024, and SAN_2024 are Rank 11 (simulated-use MRMC studies providing Technical Performance evidence per MDCG 2020-1, not primary clinical data per MDR Article 2(48)); and the legacy device market data is Rank 7 (complaints and vigilance data). In addition, 4 published peer-reviewed severity validation studies (APASI_2025, AUAS_2023, AIHS4_2023, ASCORAD_2022) provide Technical Performance evidence per MDCG 2020-1 Pillar 2 for the device's severity scoring algorithms; and NMSC_2025 provides Pillar 3 Clinical Performance evidence for malignancy detection (BCC and cSCC) in a specialist head and neck surgery setting. The sufficiency argument is:

• Coverage of indications is justified through the 3-tier evidence structure (Tier 1: malignant conditions with individual analysis; Tier 2: rare diseases with grouped analysis; Tier 3: general conditions with risk-based pooled analysis) and the 7-category epidemiological framework, demonstrating coverage of 5 of 7 categories (97% of presentations). Two acceptable gaps (autoimmune diseases 3%, genodermatoses 1%) are declared per MDCG 2020-6 § 6.5(e) with justification and linked PMCF activities.
• Demographic representativeness is demonstrated across genders, age groups (paediatric 6.3%, adult 69.5%, geriatric 19.6%), and Fitzpatrick phototypes I–V, with underrepresented phototypes identified as a PMCF monitoring priority.
• Safety is confirmed by two independent sources: zero adverse events across 800+ patients in clinical investigations and zero serious incidents across 4,500+ reports from the legacy device market experience, appraised under MDCG 2020-6 § 6.2.2.
• All predefined acceptance criteria, derived from and traceable to SotA performance benchmarks in R-TF-015-011, have been met across all three clinical benefits (7GH, 5RB, 3KX).

10. Clinical evidence strategy and regulatory methodology A dedicated section, "Clinical evidence strategy and regulatory methodology," has been added to the CER (R-TF-015-003), placed immediately following the "How to read this CER" section. This section provides, in continuous prose, a consolidated narrative of the regulatory methodology and combined evidence strategy underlying this evaluation, in accordance with MDR Article 61(1), Annex XIV Part A, MEDDEV 2.7.1 Rev 4, and MDCG 2020-1.

The section covers the following elements:

• The regulatory basis and evaluation mandate: the obligation to demonstrate conformity with GSPRs 1, 8, and 17 on the basis of sufficient clinical evidence per MDR Article 61(1), and the justification for adopting a combined strategy.
• The three complementary evidence routes with their regulatory grounding: Route A (systematic literature review, MDR Article 61(3)(a); Annex XIV Part A §1(a)); Route B (equivalence with the equivalent legacy device, MDR Articles 61(4–5); Annex XIV Part A §3; MDCG 2020-5); and Route C (own clinical investigations, MDR Article 61(4); Articles 62–82; Annex XV).
• Mapping of the combined evidence to the three MDCG 2020-1 evidence pillars: Pillar 1 (Valid Clinical Association, established through Route A); Pillar 2 (Technical Performance, the MRMC studies and four severity validation publications); and Pillar 3 (Clinical Performance, the six prospective clinical investigations, NMSC_2025 as a published Pillar 3 malignancy study, and the legacy device PMS data).
• Evidence hierarchy and weighting per MDCG 2020-6 Appendix III, with all evidence assigned to their respective ranks: Ranks 2 and 4 (prospective clinical investigations), Ranks 5–6 (retrospective validation publications), Rank 7 (legacy device PMS data), and Rank 11 (MRMC simulated-use studies).
• The risk-proportionate tiered evidence structure (Tier 1: individual per-condition analysis for malignant conditions; Tier 2: grouped analysis for rare diseases; Tier 3: pooled aggregate analysis for general conditions), with cross-reference to the detailed "Tiered evidence assessment strategy" section.
• The pre-market sufficiency determination per MDCG 2020-6 § 6.4, including the identified residual uncertainties (autoimmune diseases, genodermatoses, darker skin phototypes) and their mitigation through targeted PMCF activities per MDR Article 61(11) and Annex XIV Part B.