Adequacy review — MDCG 2020-13 Section C

Purpose and method

This report is a per-criterion adequacy review of R-TF-015-003 Clinical Evaluation Report against the MDCG 2020-13 Clinical Evaluation Assessment Report (CEAR) template, Section C (device description, classification, clinical evaluation plan, information materials supplied by the manufacturer, common specifications and harmonised standards applied, equivalence and state of the art).

The coverage matrix on the previous page already verifies structural coverage — every MDCG 2020-13 criterion has a live anchor in the CER. This review answers the complementary semantic question: for each criterion, would a notified-body reviewer tick the box in the CEAR template, or would they raise a non-conformity?

For every criterion the report records:

• the criterion verbatim from the MDCG 2020-13 template;
• the corresponding location(s) in R-TF-015-003;
• a verdict — Tick (evidence sufficient), Conditional (minor clarification required), or Fail (material gap);
• a short rationale grounded in the CER content; and
• a surgical remediation proposal for every Conditional and Fail verdict, specifying the anchor, the exact wording, and the primary source from which the wording is drawn.

Source documents used for the review:

• MDCG 2020-13 Section C (part 1) — device description and classification
• MDCG 2020-13 Section C (part 2) — CEP, information materials, equivalence, state of the art
• R-TF-015-003 Clinical Evaluation Report
• R-TF-015-001 Clinical Evaluation Plan
• R-TF-015-011 State of the Art
• Coverage matrix

Reviewer: BSI clinical-auditor reviewing agent (emulating the profile of BSI reviewers Erin Preiss and Nick). Review baseline commit: 46704ac50. Line numbers quoted below are from the CER as of that commit.

C.1 Device description and classification

C.1.1 — Device description (intended patient population, functional elements, principles of operation, novel features)

Criterion (MDCG 2020-13 §C.1): describe the device and comment on the intended purpose, including (a) the intended patient population and medical conditions; (b) the key functional elements (including software); (c) the principles of operation and mode of action, with explanation of any novel features.

CER location: #device-description — lines 326-537, 539-590.

Verdict: Conditional.

Rationale. Functional elements are well described (lines 338-372 cover the three output categories, six binary safety indicators, and the API-layer vs. physician-interface distinction) and principles of operation are explicit (lines 517-537 name the Vision Transformer architecture, object detection for counts, semantic segmentation for extent, and image-recognition for intensity). However, the intended patient population is not summarised inline at the device-description anchor — a reader must infer it from the indications/contraindications block at lines 416-440. For a Class IIb MDSW covering 346 ICD-11 categories from paediatric to geriatric and Fitzpatrick I-VI, a notified-body reviewer expects a single inline sentence. In addition, the novel features (an ICD-11-aligned probability distribution rather than a binary or short-list classifier, and the P₂=1 architectural severity constraint) are described matter-of-factly and are not flagged as novel, so a reviewer unfamiliar with comparable MDSW cannot distinguish them.

Recommended fix.

• Anchor: #device-description, immediately after the principle-of-operation paragraph (around line 537).
• Insert: "The device is intended for use in adult and paediatric patients presenting with skin findings across Fitzpatrick phototypes I-VI, in primary care, general dermatology, and specialist referral settings. Two design features are novel relative to the legacy device and to comparable MDSW: (i) the output is a complete normalised probability distribution across 346 validated ICD-11 categories rather than a binary or short-list classifier, and (ii) the architecture enforces an irreducible severity-prioritisation constraint (P₂=1) that prevents under-triage of malignant or pre-malignant findings regardless of model probability ranking."
• Primary source: R-TF-015-001 Clinical Evaluation Plan — §Intended Purpose and Target Populations; R-TF-015-003 lines 338-372 (existing output description) and 2143-2193 (P₂=1 architectural constraint, already documented).

C.1.2 — Classification (applicable rule and indents)

Criterion (MDCG 2020-13 §C.1): list the applicable classification rule(s) and indents.

CER location: #scope-of-the-clinical-evaluation — line 231; classification statements at lines 23, 289, 1204.

Verdict: Fail.

Rationale. The CER asserts Class IIb at lines 23, 289, and 1204 ("Under the MDR framework, and in accordance with the classification rules stipulated in Annex VIII, the device has been reclassified as Class IIb") but it never names Rule 11 nor the indent driving the IIb classification. For MDSW under MDR, the specific rule and indent are a CEAR Section C minimum-content item; without this, a reviewer cannot verify the classification rationale from the CER alone and would raise a non-conformity on principle.

Recommended fix.

• Anchor: #scope-of-the-clinical-evaluation, at the end of the classification paragraph.
• Insert: "The device is classified as Class IIb under MDR Annex VIII, Rule 11, second indent — software intended to provide information used to take decisions with diagnosis or therapeutic purposes, where such decisions may cause serious deterioration of a person's state of health. The Class IIb trigger is the inclusion of 13 malignant neoplasms, pre-malignant categories, and high-risk non-malignant findings within the device's probability distribution; delayed or missed identification of these presentations can lead to serious deterioration of health. Rule 11 is the sole applicable Annex VIII classification rule; no other rule yields a higher class for this device."
• Primary source: MDR 2017/745 Annex VIII Rule 11; MDCG 2019-11 (Qualification and classification of software); R-TF-015-001 Clinical Evaluation Plan §Classification.

C.1.3 — Device configurations / variants

Criterion (MDCG 2020-13 §C.1): include the manufacturer's description of the sizes, differences in design features, and different configurations; include an image of the device where possible; describe the device history and/or changes since the last assessment.

CER location: #variants — lines 456-458.

Verdict: Tick.

Rationale. The CER states "No variants" (lines 456-458). For a single-configuration MDSW this is a complete answer; the device-history element of the criterion is addressed separately under #previous-version-of-the-device (see C.1.5 below).

C.1.4 — Accessories or compatible devices

Criterion (MDCG 2020-13 §C.1): describe any accessories or compatible devices (or state "none"); identify any impact on clinical safety, performance, scope, or validity of the clinical evaluation.

CER location: #accessories-of-the-product — lines 460-467.

Verdict: Conditional.

Rationale. Lines 460-467 distinguish primary accessories (none — server-to-server API), secondary accessories (EHR/EMR systems developed independently by the user), and image-capture devices with a stated minimum of a 12 MP camera. For an image-analysis MDSW, the image-capture device is materially performance-determining across multiple parameters beyond resolution (focus, lighting, colour fidelity, dermoscopic vs. clinical modality). The CER should either state that image-capture devices are not accessories under MDR Article 2(2) or acknowledge that camera variability is absorbed upstream by the device's image-quality gate.

Recommended fix.

• Anchor: #accessories-of-the-product, at the end of the section around line 467.
• Insert: "Image-capture devices are not classified as accessories under MDR Article 2(2) because they are not specifically intended by the manufacturer to be used together with the device. Variability in capture hardware is mitigated upstream by the device's image-quality assessment subsystem (DIQA), which rejects inputs that fall outside validated technical-performance bounds before any clinical inference is produced. The full input-quality specification and DIQA acceptance thresholds are documented within the Pillar 2 (Technical Performance) evidence of the MDCG 2020-1 three-pillar framework."
• Primary source: MDR Article 2(2) accessory definition; R-TF-015-003 #three-pillar-evidence-framework-for-mdsw-mdcg-2020-1 (lines 787-797); R-TF-028-011 AI Risk Matrix documenting DIQA as the upstream input-safety gate of the defence-in-depth architecture.

C.1.5 — Previous generations and similar devices

Criterion (MDCG 2020-13 §C.1): verify that the manufacturer has provided (a) an overview of previous generations and (b) an overview of identified similar devices on the Union or international markets, including length of time on the market and sales volume.

CER location: #previous-version-of-the-device — lines 718-728; similar-device reference at line 1666; legacy market experience at line 29.

Verdict: Conditional.

Rationale. The previous-generation summary is adequate (lines 720-728: legacy device on market since 2020 under MDD Class I, transition rationale, non-clinical nature of changes) and the legacy-device market experience is quantified at line 29 (21 contracts, more than 4,500 reports, more than 500 practitioners, more than 1,000 patients, no serious incidents, no FSCA). For the five named similar devices (SkinVision, DERM, Dermalyser, ModelDerm, HUVY) at line 1666, the CER lists names only. The CEAR template expects length of time on the market and sales volume for similar devices where such information is available, so that a reviewer can grade whether the comparator set is representative of the state of the art.

Recommended fix.

• Anchor: #current-knowledge---state-of-the-art, immediately after the mention of similar devices (line 1666 area).
• Insert: "For each similar device named, the following commercialisation context was considered when reviewing user manuals and publicly available safety information: time on the EU market, regulatory class under MDD/MDR, indications coverage relative to the device under evaluation, and any publicly reported field-safety actions retrieved from MAUDE and EUDAMED. The comparator-device context table — time on market, indications, regulatory class, and field-safety actions — is maintained in R-TF-015-011 State of the Art."
• Primary source: R-TF-015-011 State of the Art — §Comparable devices / market context. Needs source verification: confirm R-TF-015-011 contains the comparator-context table before the wording is inserted in the CER. If the table is not already present in R-TF-015-011, the gap must first be closed in R-TF-015-011 and then summarised in the CER.

C.2 Clinical evaluation plan

C.2.1 — CEP summary covering the eight bullets of Annex XIV Part A §1a

Criterion (MDCG 2020-13 §C.2): briefly summarise the manufacturer's clinical evaluation plan and confirm that it meets the requirements of Annex XIV Part A §1a, highlighting areas requiring particular attention: (1) GSPRs requiring clinical data; (2) specification of intended purpose; (3) target groups with indications and contra-indications; (4) intended clinical benefits with outcome parameters; (5) methods for qualitative and quantitative safety examination; (6) parameters for acceptability of benefit-risk ratio per SotA; (7) benefit-risk issues relating to specific components (pharmaceutical, non-viable animal or human tissues); (8) clinical development plan with milestones and acceptance criteria.

CER location: #clinical-evaluation-plan — lines 239-248. Coverage-matrix status: External (full CEP in R-TF-015-001).

Verdict: Conditional.

Rationale. The CER explicitly defers the CEP summary to R-TF-015-001 (lines 239-248) and provides partial coverage scattered across the document: bullet 1 (GSPRs at lines 258-266), bullet 2 (intended purpose at lines 326-336), bullet 3 (target groups and contraindications at lines 416-440), bullet 4 (clinical benefits at lines 592-594), bullet 5 (safety methods at lines 1597-1711), and bullet 6 (acceptance criteria per SotA at lines 2026-2046). Bullet 7 is N/A (no pharmaceutical or non-viable tissue components) but is not explicitly marked N/A in the CER. Bullet 8 (clinical development plan: exploratory → confirmatory → PMCF with milestones) is not summarised in the CER — a reader cannot identify the exploratory, confirmatory, and PMCF phases or their milestones without opening the CEP. For an "External" status to earn a Tick, the CER's own summary must give a reviewer enough context to grade the external document without opening it; bullet 8 fails that test.

Recommended fix.

• Anchor: #clinical-evaluation-plan at line 239, insert a concise "CEP at a glance" subsection.
• Insert: "The full Clinical Evaluation Plan is maintained as a controlled document (R-TF-015-001) and is summarised here against Annex XIV Part A §1a: (1) GSPRs requiring clinical data — 1, 8, 17.1; (2) intended purpose — diagnostic decision support across 346 ICD-11 dermatological categories; (3) target groups, indications, contraindications — see §Contraindications and precautions; (4) clinical benefits — 7GH diagnostic accuracy, 5RB objective severity assessment, 3KX care pathway optimisation; (5) safety methods — qualitative (usability, vigilance, complaints) and quantitative (false-positive and false-negative rates per safety-critical category, residual-risk acceptability); (6) acceptance parameters per SotA — derived per benefit and per clinical domain with explicit safety margins of 3 to 23 percentage points; (7) specific components (pharmaceutical, non-viable animal or human tissues) — not applicable, the device is software-only with no biological component; (8) clinical development plan — exploratory legacy-device evidence (2020 onward) → confirmatory pre-market evidence (8 pivotal studies and 1 legacy clinical study) → PMCF activities, with acceptance criteria and milestones documented in the PMCF Plan."
• Primary source: R-TF-015-001 Clinical Evaluation Plan §Clinical Development Plan; R-TF-007-002 PMCF Plan.

C.2.2 — Clinical performance summary

Criterion (MDCG 2020-13 §C.2): summarise the clinical data to demonstrate the ability of the device to achieve its intended purpose as claimed by the manufacturer, leading to a clinical benefit for patients; describe the clinical benefits.

CER location: lines 592-594 (benefits introduction); #summary-of-clinical-benefits-achievement at lines 2066-2078; per-benefit conclusions at lines 2290-2306.

Verdict: Tick.

Rationale. The three claimed clinical benefits (7GH, 5RB, 3KX) are introduced concisely at lines 592-594; the consolidated achievement table at lines 2066-2078 maps each benefit to its acceptance criterion, observed magnitude, supporting studies, and Achieved status; and the per-benefit narrative conclusions at lines 2290-2306 close the loop. A reviewer can follow the chain claim → criterion → evidence → conclusion entirely within the CER.

C.2.3 — Safety summary

Criterion (MDCG 2020-13 §C.2): does the clinical evaluation adequately address the qualitative and quantitative aspects of clinical safety with clear reference to residual risks and undesirable side-effects? Summarise the clinical data on safety; briefly summarise any significant complaints, trends, or vigilance issues associated with earlier iterations.

CER location: lines 1599-1710 (Requirement on safety and GSPR 8); lines 2143-2210 (Risk management, residual risks, safety benchmarking against SotA); line 29 (legacy device market experience).

Verdict: Tick.

Rationale. Safety is addressed across qualitative and quantitative dimensions: presumption of conformity and summative usability validation traceable to AI-RISK-021 (lines 1599-1691), GSPR 8 side-effect acceptability (lines 1693-1710), 62 identified risks reduced to 8 residual risks under a defence-in-depth architecture with P₂=1 constraint (lines 2143-2193), MAUDE/EUDAMED safety benchmarking (lines 2195-2210), and legacy-device complaints and vigilance evidence (line 29 — zero serious incidents, zero FSCA across more than 4,500 reports over the review period). The qualitative/quantitative split, the residual-risk inventory, and the prior-iteration vigilance data are all present.

C.3 Common specifications, harmonised standards, other solutions

C.3.1 — Common specifications relevant and complied with

Criterion (MDCG 2020-13 §C.3): are there common specifications relevant to the device under evaluation? Have they been complied with? If not, explain deviations and confirm equivalent solutions per Article 9(3).

CER location: #applicable-standards-and-guidance-documents — lines 295-324.

Verdict: Conditional.

Rationale. No MDR Article 9 Common Specifications are listed, and the CER does not state whether any apply. For dermatology MDSW no Common Specifications are currently in force under MDR Article 9, so the correct answer is "none apply" — but the CER must say so explicitly. As written, a reviewer cannot tell whether Common Specifications were considered and found inapplicable, or whether they were simply overlooked.

Recommended fix.

• Anchor: #applicable-standards-and-guidance-documents, immediately before the standards list.
• Insert: "Common Specifications adopted under MDR Article 9 were screened for applicability. As of the date of this report, no Common Specifications in force under MDR Article 9 apply to the device's intended purpose, classification, or technology. This screening will be re-performed at each scheduled update of this clinical evaluation."
• Primary source: MDR 2017/745 Article 9; European Commission register of Common Specifications; R-TF-015-001 Clinical Evaluation Plan §Applicable regulatory framework.

C.3.2 — Harmonised standards relevant and applied (latest revision)

Criterion (MDCG 2020-13 §C.3): are there harmonised standards relevant to the clinical evaluation of the device? Have they been applied? If partially applied, justify and confirm equivalent level of safety and performance. Is the most up-to-date revision being used?

CER location: lines 295-324 (applicable standards list); lines 1238-1245 (preclinical standards compliance table); lines 1603-1609 (MDR harmonisation limitations for MDSW/AI and use of state-of-the-art standards).

Verdict: Fail.

Rationale. Three independent defects compound into a Fail rather than a Conditional.

1. Typographical error in the preclinical compliance table. Line 1242 lists "ISO 14791:2019" with the description "Medical devices - Application of risk management to medical devices" — ISO 14791 does not exist; the correct designation is ISO 14971:2019. On a safety-critical risk-management standard, an incorrect number in a "Full application" row will be flagged on first reading, and the same standard is cited correctly at line 319, which makes the inconsistency self-evident.
2. Second probable typographical error on a guidance document. Line 1244 lists "ISO 24791-2/2020-06" with the description "Medical devices - Guidance on the application of ISO 14971" — ISO 24791-2 is not a medical-device standard; the correct designation is ISO/TR 24971:2020.
3. Revision inconsistencies between the applicable-standards list and the preclinical compliance table. The applicable-standards list (lines 295-324) cites "EN 62304-1:2021" (line 321) while the preclinical compliance table (line 1239) cites "IEC 62304:2006/A1:2015"; ISO/IEC 62366-1:2015 at line 322 vs. ISO 62366-1:2015/A1:2020 at line 1245 — the amendment appears in the table but not in the list. The two sources should be reconciled so that a reviewer can confirm the most up-to-date revision is in use.
4. No harmonisation-status label per standard. Neither the list nor the table indicates, per row, whether each standard is harmonised under MDR (with OJEU publication reference), harmonised under MDD only and applied as state of the art under MDR per MDCG 2021-5, or a non-harmonised consensus standard. Presumption of conformity attaches only to MDR-harmonised standards, so this label is material.

Recommended fix.

• Anchor 1: line 1242 — correct "ISO 14791:2019" to "ISO 14971:2019".
• Anchor 2: line 1244 — correct "ISO 24791-2/2020-06" to "ISO/TR 24971:2020".
• Anchor 3: reconcile the applicable-standards list (lines 295-324) with the preclinical compliance table (lines 1238-1245) so the revision cited for each standard is identical in both places and is the most up-to-date revision applied by the organisation.
• Anchor 4: #applicable-standards-and-guidance-documents — add a status column or per-row footnote indicating for each standard whether it is "Harmonised under MDR (OJEU [reference, date])", "Harmonised under MDD only — applied as state of the art under MDR per MDCG 2021-5", or "Non-harmonised consensus standard — applied as state of the art".
• Insert (after the list): "Where a standard is listed as harmonised under MDD only, it has been applied as state of the art under MDR pending publication of an MDR-harmonised equivalent. This approach is consistent with transitional guidance in MDCG 2021-5."
• Primary source: OJEU consolidated list of harmonised standards under Regulation (EU) 2017/745; MDCG 2021-5; R-TF-001-005 List of applicable standards and regulations; R-TF-015-001 Clinical Evaluation Plan §Standards strategy.

C.3.3 — Other solutions applied

Criterion (MDCG 2020-13 §C.3): describe any other standards, guidance, or solutions that have been applied, and the manufacturer's justification.

CER location: lines 295-324 (MDCG series, IMDRF N43/N56/N57/N65); lines 767-773 (three methodological frameworks — MDCG 2020-6, MEDDEV 2.7.1 Rev 4, MDCG 2020-1); lines 1603-1611 (GMLP and AI Act principles applied to AI performance testing).

Verdict: Tick.

Rationale. Non-harmonised guidance is listed and the rationale for using each is given: lines 1603-1611 explain the MDSW/AI guidance gap and the use of Good Machine Learning Practice and AI Act principles; lines 767-773 explicitly identify the three methodological frameworks (MDCG 2020-6, MEDDEV 2.7.1 Rev 4, MDCG 2020-1) as the structuring guidance for the clinical evaluation.

C.4 Demonstration of equivalence

C.4.1 — CE based on clinical investigations / other studies of an equivalent device

Criterion (MDCG 2020-13 §C): is the clinical evaluation based upon clinical investigation(s) or other studies reported in scientific literature, of a device for which equivalence to the device in question can be demonstrated? (Yes/No + device named.)

CER location: line 27; lines 1116-1229 (full equivalence demonstration); line 1183 (same-manufacturer access); line 1196 (applicability of legacy data).

Verdict: Tick.

Rationale. The equivalence claim is unambiguous: equivalence is asserted only to the manufacturer's own legacy device, the equivalent device is named at line 1183, and the rationale (same manufacturer, full access to design, technical documentation, and performance data) is stated. Lines 27 and 1196 confirm that legacy-device clinical data are used as supporting evidence on the basis of demonstrated equivalence.

C.4.2 — CE based on peer-reviewed literature on an equivalent device

Criterion (MDCG 2020-13 §C): is the clinical evaluation based upon reports published in peer-reviewed scientific literature on a device for which equivalence can be demonstrated? (Yes/No + source + most relevant device.)

CER location: lines 1014-1050 (published severity-validation studies of the device itself); lines 1353-1471 (systematic literature search on the device under evaluation); lines 1664-1666 (similar devices reviewed for safety alignment, not equivalence).

Verdict: Tick.

Rationale. Equivalence to a third-party device via peer-reviewed literature is not claimed, and the CER is clear about this distinction: the published severity-validation studies at lines 1014-1050 are validations of the device itself (Pillar 2 Technical Performance evidence), the systematic literature search at lines 1353-1471 provides SotA context rather than equivalent-device evidence, and the five named similar devices at lines 1664-1666 are reviewed for safety alignment rather than for equivalence. The negative answer is well-supported.

C.4.3 — Assessment of equivalence (rationales, exclusion of non-equivalent data, three criteria)

Criterion (MDCG 2020-13 §C): indicate which devices are/are not equivalent; confirm that data from non-equivalent devices are excluded; each equivalence demonstration based on a single device meeting all three criteria (clinical, technical, biological).

CER location: #demonstration-of-equivalence — lines 1116-1183; biological N/A rationale at lines 1175-1179.

Verdict: Tick.

Rationale. A single equivalent device is identified (the legacy device); all three Annex XIV §3 criteria (technical, clinical, biological) are addressed; biological is explicitly marked not applicable with stated rationale (software-only, no tissue or fluid contact) at lines 1175-1179. One demonstration, one equivalent device, no data from non-equivalent devices incorporated as equivalence evidence.

C.4.4 — Devices equivalent per Annex XIV §3 (differences justified)

Criterion (MDCG 2020-13 §C): are the devices equivalent per Annex XIV §3 (technical, biological, clinical)? Identify any differences and verify why these are not expected to adversely affect safety and performance.

CER location: lines 1120-1156 (technical equivalence table with the four MDR-alignment changes enumerated); lines 1158-1173 (clinical equivalence table); lines 1175-1179 (biological N/A); lines 1181-1196 (conclusions).

Verdict: Conditional.

Rationale. Technical equivalence at lines 1120-1156 lists four MDR-alignment changes (migration to microservices architecture, HL7 FHIR adoption, database encryption, enhanced UI feedback) and asserts they do not affect clinical safety or performance. Clinical equivalence at lines 1158-1173 and biological N/A at lines 1175-1179 are well handled. For each of the four technical differences, however, the CER provides a global non-affect statement rather than a per-change clinical-relevance assessment. A reviewer will expect each enumerated difference to be mapped to an explicit clinical-relevance conclusion.

Additional observation flagged on independent second-pass review: the Critical Performance Requirements row at line 1151 asserts "AUC ≥ 0.9 for malignancy, specificity ≥ 80%, sensitivity ≥ 75% (validated in studies)" and marks the legacy device "Same" without citing the legacy-device validation evidence on which the "Same" assertion rests. Whether the same numerical thresholds were demonstrated for the legacy device (and not merely claimed) is the basis on which the technical-performance row is auditable, so an evidence pointer is needed.

Recommended fix.

• Anchor 1: insert a short per-change clinical-relevance table immediately after the four-change enumeration at line 1190, before the "Justification for Lack of Clinical Impact" subheading at line 1192.
• Insert (table): a four-row table with columns "Change", "Mechanism", "Clinical-relevance assessment", "Supporting evidence" — one row each for (1) microservices migration (backend architecture; inference pipeline and model weights unchanged; output bit-equivalence verified; no impact on clinical output; supporting evidence = regression V&V records in the technical documentation); (2) HL7 FHIR adoption (output serialisation and EHR interchange; report content rendered to clinician identical to the legacy device; supporting evidence = interoperability V&V and IFU validation); (3) database encryption and cybersecurity upgrades (data-at-rest and in-transit protection; no effect on inference output; supporting evidence = cybersecurity V&V and R-TF-013-002 patient-safety risk register); (4) enhanced UI feedback (user-facing image-quality prompts; mitigates usability risk AI-RISK-021; validated post-change in the October 2025 summative usability study; supporting evidence = R-TF-025-007 Summative Evaluation Report). Retain the existing global statement at line 1194 as the closing summary.
• Anchor 2: the Critical Performance Requirements row at line 1151 — add a cell-level citation to the legacy-device validation evidence that supports the "Same" designation (for example, the legacy-device performance validation records held in the MDD Technical File).
• Primary source: R-TF-015-003 lines 1120-1156 (existing technical comparison); AI-RISK-021 summative usability validation in R-TF-025-007; R-TF-028-011 AI Risk Matrix documenting the P₂=1 invariant; legacy-device MDD Technical File performance records (cross-reference to be added).

C.5 Access to data

C.5.1 — Manufacturer's access to data for equivalent devices

Criterion (MDCG 2020-13 §C): comment on the manufacturer's access to data relating to equivalent devices sufficient to justify equivalence claims.

CER location: line 1183.

Verdict: Tick.

Rationale. "Both products were developed by the same manufacturer... there is full access to the design, technical documentation, and performance data of both devices" — this is the strongest form of access statement and exactly what the CEAR template seeks.

C.5.2 — Article 61(5) contract (implantable / Class III requirement)

Criterion (MDCG 2020-13 §C): for implantable and Class III devices, if equivalence is claimed with a device marketed by another manufacturer, confirm that there is a current valid contract between the two manufacturers allowing ongoing access to the technical documentation in accordance with Article 61(5).

CER location: not explicitly addressed.

Verdict: Conditional.

Rationale. Article 61(5) applies only when equivalence is claimed to another manufacturer's device, and is mandatory only for implantable and Class III devices. For this Class IIb non-implantable device with same-manufacturer equivalence, the clause does not apply — but the CEAR template expects an explicit acknowledgement, not silence.

Recommended fix.

• Anchor: #conclusions-regarding-equivalence (line 1181), insert one sentence.
• Insert: "MDR Article 61(5) (contractual access to technical documentation of an equivalent device manufactured by a third party) does not apply: equivalence is claimed to the manufacturer's own legacy device, and the device under evaluation is Class IIb non-implantable. Article 61(5) is triggered only for implantable and Class III devices when equivalence is to a third-party device."
• Primary source: MDR 2017/745 Article 61(5); same-manufacturer status at R-TF-015-003 line 1183.

C.5.3 — Original CE of equivalent device performed under MDR

Criterion (MDCG 2020-13 §C): has the original clinical evaluation been performed in compliance with the requirements of Regulation 2017/745, and has the manufacturer of the second device provided clear evidence thereof?

CER location: lines 720-728 (legacy regulatory status); lines 1200-1204 (class transition); line 1207 (sufficiency statement); lines 1224-1228 (standalone MDR Technical Documentation).

Verdict: Conditional.

Rationale. The CER is honest that the legacy device was CE-marked under MDD as Class I (lines 720-728, 1200-1204) and explicitly acknowledges that "the clinical documentation and data compiled for the Class I legacy device are not, in themselves, sufficient to satisfy the level of scrutiny required for a Class IIb device" at line 1207. It also notes that a standalone MDR Technical Documentation file has been prepared (lines 1224-1228). The CEAR §C.5.3 question, however, is specifically whether the clinical evaluation of the equivalent device was performed under MDR — the answer is no, but the CER does not put that sentence directly next to the equivalence conclusion. The link between the honest acknowledgement and the resolution (additional pre-market clinical evidence generated specifically for the device under evaluation) should be made explicit.

Recommended fix.

• Anchor: #demonstration-of-equivalence, immediately before line 1207.
• Insert: "The equivalent device's original CE marking and clinical evaluation were performed under MDD (Class I), not under MDR 2017/745. Legacy clinical evidence is therefore used only as supporting evidence within the MDR clinical evaluation of the device under evaluation; the MDR-level evidence requirement is met by the pre-market clinical investigations and the post-market real-world evidence generated specifically for the device under evaluation, listed in §Pre-market clinical investigations and §Clinical data generated from risk management and PMS activities."
• Primary source: R-TF-015-003 lines 720-728 (legacy regulatory status); R-TF-015-001 Clinical Evaluation Plan §Evidence strategy.

C.5.4 — Access to data sufficient to support equivalence

Criterion (MDCG 2020-13 §C): confirm that access to data is sufficient to provide enough information about the equivalent device to support equivalence claims, including any testing undertaken to confirm equivalence of specifications or performance.

CER location: lines 1120-1156 (technical specification comparison); line 1183 (same-manufacturer access); line 1207 (acknowledged limits of legacy data).

Verdict: Tick.

Rationale. Same-manufacturer access is documented at line 1183, the technical specification comparison is detailed at lines 1120-1156, and the CER explicitly acknowledges at line 1207 that the legacy data are not in themselves sufficient and so additional pre-market evidence has been generated. Access is sufficient and the limits of legacy evidence are honestly bounded.

C.5.5 — Other limitations with respect to equivalent devices

Criterion (MDCG 2020-13 §C): comment on any other limitations with respect to the equivalent devices or the manufacturer's equivalence claims, and the extent to which these limitations impact the clinical evaluation and its conclusions.

CER location: lines 1198-1223 (Justification for Additional Clinical Evidence).

Verdict: Tick.

Rationale. The section explicitly enumerates the limitations of the legacy MDD evidence base (class change MDD Class I → MDR Class IIb, GSPRs replacing Essential Requirements, MDR-specific requirements on clinical validation, risk management, usability, and software validation per GSPR 17). The evidence gap is named at line 1214 and the closure approach specified (additional pre-market clinical investigations). This is the level of limitation-acknowledgement the CEAR seeks.

C.6 State of the art

C.6.1 — Alternative treatment options with comparable safety / performance

Criterion (MDCG 2020-13 §C): describe the alternative available treatment options identified by the manufacturer that could offer comparable safety and performance for the same treatment indications and patient populations.

CER location: #current-knowledge---state-of-the-art — lines 730-755; similar MDSW at lines 1664-1666. Full SotA in R-TF-015-011 (External per coverage matrix).

Verdict: Conditional.

Rationale. The CER summarises SotA for specialist dermatology access, PCP-vs-dermatologist performance gaps, AI adjunctive use, and teledermatology at lines 730-755, and lists similar MDSW devices at lines 1664-1666. "Alternative treatment options" in the CEAR sense includes both alternative diagnostic pathways (unaided clinical examination, dermoscopy, teledermatology, biopsy as ground truth) and alternative diagnostic-decision-support technologies (other AI MDSW, total-body photography, reflectance confocal microscopy). The CER summary leans on the latter and underspecifies the former. An External SotA status requires the CER summary to enumerate the pathways considered so a reviewer can grade scope without opening R-TF-015-011.

Recommended fix.

• Anchor: #current-knowledge---state-of-the-art, after the SotA data-source table (around line 755).
• Insert: "Alternative diagnostic pathways considered as state-of-the-art comparators include: (i) unaided clinical examination by general practitioners; (ii) unaided clinical examination by general dermatologists; (iii) dermoscopic examination by trained users; (iv) teledermatology referral with asynchronous specialist review; (v) other AI-based diagnostic decision-support MDSW (see §Similar devices); and (vi) reference-standard pathways (biopsy with histopathology). The full comparative analysis of safety and performance for each pathway is documented in R-TF-015-011."
• Primary source: R-TF-015-011 State of the Art §Alternative pathways. Needs source verification: confirm R-TF-015-011 covers these six pathways before the wording is inserted in the CER. If R-TF-015-011 does not, the gap must first be closed in R-TF-015-011.

C.6.2 — Benchmarks from aggregate data or individual devices with rationale

Criterion (MDCG 2020-13 §C): briefly describe how benchmarks for safety and performance have been identified in terms of the state of the art. Benchmarks normally based on aggregate data (systematic reviews or registry analysis); if individual devices are benchmarks, provide a rationale.

CER location: lines 746-755 (PRISMA systematic search, 227 candidates → 64 included, quality/relevance/level-of-evidence metrics); lines 2032-2046 (per-domain meta-analytic baselines); lines 2195-2210 (MAUDE/EUDAMED safety benchmarking).

Verdict: Tick.

Rationale. Benchmarks are derived from aggregate data (PRISMA systematic search with weighted appraisal), meta-analytic baselines are explicit per clinical domain at lines 2032-2046 (melanoma, multiple malignant conditions pooled, diagnostic accuracy improvement, IHS4 ICC, alopecia Kappa), and safety benchmarks come from MAUDE/EUDAMED vigilance registry analysis at lines 2195-2210. The methodology and provenance of each benchmark are traceable.

C.6.3 — SotA based on appropriate literature search

Criterion (MDCG 2020-13 §C): confirm that the manufacturer's description of the state of the art is based upon an appropriate literature search (see Section D).

CER location: lines 746-755 (SotA literature search summary); lines 1353-1471 (device-related literature search); lines 1982-1995 (April 2026 supplementary literature review).

Verdict: Tick.

Rationale. Two complementary searches are documented and cross-referenced: the SotA literature search (PRISMA, 227 candidates → 64 included, with quality, relevance, and level-of-evidence metrics) and the device-related literature search at lines 1353-1471. The April 2026 Supplementary Literature Review (lines 1982-1995) closed eight specific evidence gaps with 25 additional papers, bringing the corpus to 93 appraised papers. The chain from search protocol → execution → appraisal → corpus is complete.

C.6.4 — Previously marketed devices: SotA still accurate, device still state of the art

Criterion (MDCG 2020-13 §C): for devices previously marketed, is the description of the state of the art still accurate? Can the device still be considered to be state of the art?

CER location: lines 1982-1995 (Supplementary Literature Review); line 29 (legacy market experience); lines 714-716 ("this product has not been commercialized yet").

Verdict: Conditional.

Rationale. The Supplementary Literature Review explicitly updates the SotA for nine targeted areas and confirms the device remains aligned with current state of the art. However, line 714-716 ("this product has not been commercialized yet") creates an apparent inconsistency with the legacy device's market history at line 29 that a reviewer will flag. The C.6.4 question turns on whether the equivalent legacy device counts as "previously marketed" for the purpose of triggering the SotA-still-accurate test — and the CER does not resolve that explicitly.

Recommended fix.

• Anchor: #current-knowledge---state-of-the-art, within the SotA-update narrative (around line 1982).
• Insert: "Although the device under evaluation has not yet been placed on the market, the equivalent legacy device has been on the market since 2020 and has accumulated over 4,500 reports across more than 500 practitioners. For the purposes of MDCG 2020-13 §C.6.4, the legacy device's market history is treated as previously-marketed experience. The state of the art has been re-confirmed as current via the April 2026 Supplementary Literature Review, and the device under evaluation remains consistent with current state of the art across all evaluated clinical domains."
• Primary source: R-TF-015-003 line 29 (legacy market experience); R-TF-015-003 lines 1982-1995 (Supplementary Literature Review).

C.6.5 — Safety / performance endpoints appropriate, clinically relevant, surrogates justified

Criterion (MDCG 2020-13 §C): what performance and safety endpoints has the manufacturer identified? In light of outcomes achievable with benchmarks, are these endpoints appropriate and clinically relevant? Have surrogate endpoints been adequately justified?

CER location: lines 539-590 (metric glossary and definitions); lines 541-557 (metric-to-SotA traceability table); lines 2050-2065 (Bayes' theorem predictive-value analysis by clinical setting per MEDDEV 2.7.1 Rev 4 Annex A7.3).

Verdict: Tick.

Rationale. Each endpoint (Top-K accuracy, AUC, sensitivity, specificity, PPV, NPV, ICC, Cohen's Kappa, expert consensus, efficiency metrics) is defined in a glossary and traced to peer-reviewed SotA literature (lines 541-557). The Bayes' theorem analysis at lines 2050-2065 across primary care, general dermatology, and pigmented-lesion-clinic settings (per MEDDEV 2.7.1 Rev 4 Annex A7.3) is exactly the kind of setting-specific predictive-value translation that confirms clinical relevance. Surrogate endpoints (Top-K accuracy as a surrogate for clinical-decision quality) are implicitly justified by the SotA-traceability table.

C.6.6 — Indicative parameters for acceptability of benefit-risk ratio per SotA

Criterion (MDCG 2020-13 §C): has the manufacturer adequately described an indicative list and specification of parameters used to determine, based on the state of the art in medicine, the acceptability of the benefit-risk ratio for the various indications and for the intended purpose or purposes of the device?

CER location: lines 2016-2024 (three-stage derivation rationale); lines 2026-2046 (per-benefit × per-domain acceptance-criteria derivation table); lines 2116-2193 (benefit-risk profile assessment, defence-in-depth, P₂=1).

Verdict: Tick.

Rationale. The Acceptance Criteria Derivation table at lines 2026-2046 is structured exactly as MDCG 2020-13 §C.6.6 expects: per benefit ID × clinical domain, with relevant SotA article(s), methodology (meta-analysis, weighted average, literature-benchmarked range, or direct observation), derived SotA baseline, and acceptance criterion with explicit safety margin. The three-stage rationale at lines 2016-2024 (extraction of SotA benchmarks → synthesis of baselines → establishment of targets with domain-specific safety margins of 3 to 23 percentage points) is explicit. The benefit-risk profile assessment at lines 2116-2193 (defence-in-depth, P₂=1, residual-risk acceptability) closes the loop.

Section C overall

The table below records the roll-up verdict for each of the six MDCG 2020-13 Section C compliance tick-boxes.

§C tick-box	Verdict	One-sentence rationale
Device details, intended purpose, classification	Compliant with minor non-compliance	Classification fails to name MDR Annex VIII Rule 11 (C.1.2 Fail), and device description + accessories need explicit intended-population and novel-features statements (C.1.1 and C.1.4 Conditional); all surgical fixes available.
Clinical evaluation plan	Compliant with minor non-compliance	CEP is appropriately externalised to R-TF-015-001, but the CER summary omits Annex XIV Part A bullet 8 (clinical development plan phases and milestones) and does not mark bullet 7 N/A (C.2.1 Conditional); C.2.2 and C.2.3 are Tick.
Common specifications, harmonised standards, other solutions	Non-compliant	Two standard-designation typographical errors (ISO 14791/ISO 14971 at line 1242; ISO 24791-2/ISO/TR 24971 at line 1244), revision inconsistencies between the standards list and the preclinical compliance table, and no MDR-vs-MDD harmonisation status label per standard together escalate C.3.2 to Fail; C.3.1 remains Conditional (no explicit "no Common Specifications apply" statement); C.3.3 is Tick.
Demonstration of equivalence	Compliant with minor non-compliance	Equivalence framework is sound (C.4.1 – C.4.3 Tick), but each of the four MDR-alignment technical changes needs a per-row clinical-relevance assessment rather than a global non-affect statement (C.4.4 Conditional).
Equivalence — access to data	Compliant with minor non-compliance	Same-manufacturer access is unambiguous (C.5.1, C.5.4, C.5.5 Tick), but the CER must explicitly acknowledge that Article 61(5) does not apply (C.5.2 Conditional) and that the equivalent device's CE was under MDD, not MDR (C.5.3).
State of the art	Compliant with minor non-compliance	SotA benchmarks, search methodology, endpoint appropriateness, and benefit-risk acceptance parameters are Tick (C.6.2, C.6.3, C.6.5, C.6.6); alternative-pathway enumeration (C.6.1) and previously-marketed framing (C.6.4) Conditional.

Roll-up. Section C is Compliant with minor non-compliance across five of the six tick-boxes and Non-compliant on the common-specifications / harmonised-standards tick-box. The review produced 2 Fails (C.1.2 classification rule not named; C.3.2 harmonised-standards designation errors and missing harmonisation status), 10 Conditionals, and 14 Ticks across 26 criteria, with two Conditionals requiring source verification in R-TF-015-011 before the proposed wording is inserted in the CER (C.1.5 similar-device market context, C.6.1 alternative-pathway enumeration). None of the findings represent evidence-base failures; all are wording, signposting, standard-designation, or classification-rule fixes that a notified-body reviewer would resolve via documented remediation rather than by questioning the clinical evidence itself, provided the surgical fixes are implemented before submission.

Second-pass review

Three items from the first pass were re-judged by an independent second-pass reviewer with no sight of the first-pass verdicts:

• C.1.2 (Fail) — both passes agree. Independent review confirms Fail; Rule 11 and the second indent must be named.
• C.3.2 (originally Conditional, upgraded to Fail). Independent review identified a second standard-designation error (ISO 24791-2 at line 1244 should read ISO/TR 24971:2020) and a revision inconsistency between the applicable-standards list and the preclinical compliance table, both of which the first pass had not captured. The verdict and the recommended fix have been updated accordingly above.
• C.4.4 (Conditional) — both passes agree, with the independent review adding a per-change clinical-relevance table format and a legacy-device critical-performance evidence pointer. The recommended fix above has been enriched with both additions.

The second-pass review increased the severity of one verdict and enriched one fix; no verdict was downgraded.

Next steps

Adequacy reviews for the remaining MDCG 2020-13 sections will be produced in dedicated files in this directory once the Section C findings have been agreed:

• Section D — Literature review
• Section E — Clinical investigations
• Section F — PMS / PMCF
• Section G — IFU, SSCP, labelling
• Overall conclusions
• Sections A / B — Administrative and reviewer roster
• Sections I / J / K — Article 54, Article 61(10), and Article 61(2) considerations (all N/A for this device, but explicit N/A rationale still required)

Once all sections are reviewed, a consolidated summary table will be added at adequacy-review/index.mdx and linked from the round-1 landing page.