R-TF-015-004 Clinical investigation plan

Scope

This Clinical Investigation Plan (CIP) sets out the rationale, objectives, design, methodology, conduct, implementation, record-keeping, and method of analysis for the clinical investigation.

CIP Identification

	CIP
Title of the clinical investigation	Multi-Reader Multi-Case Study for Evaluating the Diagnostic Performance of Healthcare Professionals Assisted by Legit.Health Plus on Fitzpatrick Phototype V–VI Skin Presentations
Device under investigation	Legit.Health Plus
Protocol version	Version 1.0
Date	2026-04-14
Protocol code	LEGIT.HEALTH_MAN_2025
Sponsor	AI Labs Group S.L.
Coordinating Investigator	Dr. Antonio Martorell Calatayud
Principal Investigator(s)	Dr. Antonio Martorell Calatayud
Investigational site(s)	This study is conducted remotely through a centralized web-based platform.
Ethics Committee	This study does not require Ethics Committee approval because it is observational and non-interventional. All data used consists of fully anonymized images sourced from public dermatology atlases and databases, containing no information permitting patient identification. As such, the research meets the criteria for exemption from ethics committee review under applicable regulatory frameworks.

Table of contents

• Scope
• CIP Identification
• Regulatory classification of this study
• Trial registration
• Compliance Statement
• Abbreviations and definitions
• CIP or protocol specifications
  • Principal Investigator
  • Coordinating investigator
  • Collaborating Investigator(s)
  • Technical Support (AI Labs Group S.L.)
  • Investigational sites
  • Funding
• Product Identification and Description
• Justification of the design
  • Background and rationale
  • Risks and benefits of the product in investigation and clinical research
• Hypothesis
• Objectives
  • Primary objective
  • Secondary objectives
• Summary of the study
• Design and methods
  • Type of clinical research
  • Reference standard (ground truth)
  • Population
  • Sample size
    • Power calculation
    • Justification for the image set
    • Selection of 5 dermatologists
  • Dermatological conditions and case distribution
  • Duration
  • Acceptance criteria
  • Inclusion criteria
  • Exclusion criteria
  • Variables
    • Main variable
    • Secondary variables
  • Condition of interest
  • Limitations of clinical research
• Ethical considerations
  • Data confidentiality
  • Bias minimization measures
  • Calendar
  • Monitoring plan
  • Completion of the investigation
  • Statistical analysis
    • Primary analysis (MRMC methodology)
    • Secondary analyses
  • Data management
    • Data collection
    • Access control and traceability
    • Data anonymity
    • Data export and analysis
    • Data security
• CIP Modification
• CIP Deviations
• Start, follow-up and end reports
• Statements of compliance
• Informed Consent process
  • For patients/image subjects
  • For healthcare practitioners (dermatologists)
• Adverse events, adverse product reactions and product deficiencies
  • Adverse Events (AE) and Adverse Events to the Product (AEP)
  • Product deficiencies
  • Serious Adverse Events, serious adverse events to the product and serious and unexpected adverse events to the product
  • Foreseeable adverse events and adverse events to the product
  • Data Monitoring Committee (DMC)
• Suspension or early termination of clinical research

Regulatory classification of this study

This study does not constitute a clinical investigation under MDR Article 2(45), which defines a clinical investigation as "any systematic investigation involving one or more human subjects, undertaken to assess the safety or performance of a device." In this study:

• There are no human subjects (patients). The clinical images are fully anonymized photographs sourced from public dermatological atlases. No patients are recruited, examined, or affected in any way by this study.
• The participating healthcare professionals are evaluators, not research subjects. They provide diagnostic assessments in a controlled professional capacity and are not exposed to any intervention or risk.
• The study is observational and non-interventional. No clinical decisions, treatments, or patient care pathways are influenced by the study.

Consequently, the requirements of MDR Article 62 (clinical investigations), Articles 73–77 (sponsor obligations), and Annex XV (clinical investigations) do not apply. In particular, competent authority notification (AEMPS), EUDAMED registration, and formal Ethics Committee opinion are not required.

This CIP is prepared in accordance with ISO 14155:2021 as a best-practice methodological framework to ensure scientific rigour and traceability, even though its full regulatory scope is not triggered by this study design. The study results are intended as supporting clinical evidence under MDR Article 61 and MDCG 2020-6.

Trial registration

This study is not registered in ClinicalTrials.gov or the EMA RWD Catalogue (EUPAS). Registration is not required because this study does not meet the MDR definition of a clinical investigation (see "Regulatory classification of this study" above). No patients are involved, no clinical interventions are performed, and no clinical decisions are influenced by the study. The study may be voluntarily registered post-hoc if deemed appropriate for transparency purposes.

Compliance Statement

This study is conducted in accordance with the following standards and regulations, to the extent applicable to a non-interventional, observational MRMC study with no patient involvement:

• Harmonised standard UNE-EN ISO 14155:2021 (applied as best-practice methodological framework).
• Regulation (EU) 2017/745 on medical devices (MDR) — Article 61 (clinical evaluation) and Annex XIV (clinical evaluation and PMCF).
• Harmonised standard UNE-EN ISO 13485:2016.
• Regulation (EU) 2016/679 (GDPR).
• Spanish Organic Law 3/2018 on the Protection of Personal Data and guarantee of digital rights.

Since this study does not constitute a clinical investigation under MDR Article 2(45), the requirements of MDR Article 62, Articles 73–77, and Annex XV are not applicable. Ethics Committee opinion and competent authority notification are not required (see "Regulatory classification of this study").

Abbreviations and definitions

• AE: Adverse Event
• AEMPS: Spanish Agency of Medicines and Medical Devices
• AEP: Adverse Reaction to Product
• AUC: Area Under the ROC Curve
• CAD: Computer-Aided Diagnosis
• CMD: Data Monitoring Committee
• CIP: Clinical Investigation Plan
• CUS: Clinical Utility Questionnaire
• DLQI: Dermatology Quality of Life Index
• GCP: Standards of Good Clinical Practice
• ICH: International Conference of Harmonization
• IFU: Instructions For Use
• IRB: Institutional Review Board
• N/A: Not Applicable
• NCA: National Competent Authority
• PI: Principal Investigator
• PPV: Positive Predictive Value
• NPV: Negative Predictive Value
• SAE: Serious Adverse Events
• SAEP: Serious Adverse Event to Product
• SUAEP: Serious and Unexpected Adverse Event to the Product
• SUS: System Usability Scale

CIP or protocol specifications

Principal Investigator

• Dr. Antonio Martorell Calatayud

Coordinating investigator

• Dr. Antonio Martorell Calatayud

Collaborating Investigator(s)

To be confirmed upon recruitment. A minimum of 5 dermatologists will be recruited for this study.

Technical Support (AI Labs Group S.L.)

• Mr. Alfonso Medela
• Mr. Taig Mac Carthy

Investigational sites

This study is conducted remotely through a centralized web-based platform. Healthcare professionals are provided with individual user credentials (username and password) to securely access the platform. All images are presented through this platform, and practitioners' assessments are recorded on the same system. Access logs are maintained to ensure traceability of all practitioner interactions with the platform.

Funding

This research is carried out without any funding or sponsorship.

Product Identification and Description

	Information
Device name	Legit.Health Plus (hereinafter, the device)
Model and type	NA
Version	1.1.0.0
Basic UDI-DI	8437025550LegitCADx6X
Certificate number (if available)	MDR 000000 (Pending)
EMDN code(s)	Z12040192 (General medicine diagnosis and monitoring instruments - Medical device software)
GMDN code	65975
EU MDR 2017/745	Class IIb
EU MDR Classification rule	Rule 11
Novel product (True/False)	TRUE
Novel related clinical procedure (True/False)	TRUE
SRN	ES-MF-000025345

The specific device version under evaluation is version 1.1.0.0. If the device software is updated during the study period, the impact on study validity will be assessed by the sponsor, and a protocol deviation will be documented if the update affects diagnostic output. If the update is deemed to materially affect the primary endpoint, the affected cases will be re-evaluated or excluded, and the CIR will document the device version(s) used for each assessment phase.

Justification of the design

Background and rationale

Dermatological conditions represent a significant portion of primary care consultations, constituting approximately 5% of all visits. However, discrepancies between diagnoses made by general practitioners and dermatologists remain substantial, with concordance rates between 57% and 65.52%. This gap in expertise often leads to misdiagnoses, incorrect referrals, and delays in appropriate treatment. The limited availability of dermatologists, especially in rural areas, further complicates patient care, underscoring the need for innovative solutions to optimize resource allocation and improve diagnostic accuracy.

Artificial intelligence presents a transformative opportunity to enhance the diagnostic capabilities of healthcare professionals. The device has already been validated for the diagnosis of skin conditions across multiple clinical investigations (SAN_2024, BI_2024, PH_2024), demonstrating statistically significant improvements in diagnostic accuracy when healthcare professionals are assisted by the device.

However, the existing clinical evidence base is predominantly composed of clinical images representing Fitzpatrick phototypes I–IV. Fitzpatrick phototypes V and VI — corresponding to darker skin tones — are underrepresented, accounting for approximately 5% and 0% of the total study population, respectively. This gap limits the generalizability of the existing evidence to populations with darker skin tones. This study is specifically designed to address this gap by evaluating the device's diagnostic support performance on clinical images representing Fitzpatrick phototype V–VI skin presentations.

The 149 clinical images used in this study are derived from the image sets of three completed MRMC studies (SAN_2024, BI_2024, PH_2024), which collectively contained 158 unique clinical cases sourced from public dermatological atlases. The images represent Fitzpatrick phototype V–VI presentations of the same dermatological conditions already evaluated in the source studies, ensuring comparability of results across phototype groups.

Risks and benefits of the product in investigation and clinical research

In this study, there is no patient recruitment or active patient involvement. The images used are completely anonymized clinical images sourced from public dermatological atlases and freely available public sources. These images are not derived from identifiable patients, and their anonymization makes patient recognition impossible. Using the device could optimize diagnostic accuracy in clinical practice, potentially save consultation time and costs, and support better clinical decision-making for patients with skin pathologies across all Fitzpatrick phototypes. The participating HCPs will sign a contract with the sponsor to regulate their participation in the study.

Hypothesis

The information provided by the device increases the true accuracy of healthcare professionals (HCPs) in the diagnosis of multiple dermatological conditions presented on Fitzpatrick phototype V–VI skin.

Objectives

Primary objective

To validate that the information provided by the device increases the true accuracy of healthcare professionals (HCPs) in the diagnosis of multiple dermatological conditions presented on Fitzpatrick phototype V–VI skin.

Secondary objectives

• To validate the percentage of cases that should be referred according to the HCP with the information provided by the device.
• To confirm that the use of the device is perceived by specialists as being of great clinical utility across skin phototypes.

Summary of the study

This is a prospective, observational multi-reader, multi-case (MRMC) self-controlled study. It is designed to assess whether the use of the device by dermatologists can increase the accuracy in the diagnosis of multiple dermatological conditions on Fitzpatrick phototype V–VI skin. A minimum of 5 dermatologists will be presented with 149 images of different skin conditions representing dark skin phototypes. Data collection will include diagnosis accuracy for different dermatological pathologies. The study adheres to strict ethical guidelines, ensuring data confidentiality and compliance with international standards.

Design and methods

Type of clinical research

This is a prospective, observational, multi-reader, multi-case (MRMC) self-controlled study to evaluate whether the use of the medical device by dermatologists helps to increase the accuracy in the diagnosis of different skin conditions presented on Fitzpatrick phototype V–VI skin. The study uses a progressive information disclosure design: for each clinical case, the reader completes a sequence of three assessment stages that reveal progressively more device output. This within-case, within-reader comparison isolates the device's incremental contribution to diagnostic accuracy.

The three stages per case are:

1. Unassisted diagnosis (Stage 1): The reader views the clinical image and patient anamnesis, and provides their primary diagnosis without any device output.
2. Assisted diagnosis (Stage 2): The device's differential diagnosis (ICD-11 probability distribution) is additionally displayed. The reader provides their revised diagnosis.
3. Referral assessment (Stage 3): The device's malignancy probability, referral recommendation, and diagnostic entropy are additionally displayed. The reader decides whether to refer the patient.

The stages are completed sequentially for each case before the reader proceeds to the next case. The order of case presentation is independently randomised for each reader to prevent order effects. The self-controlled comparison between Stage 1 (unassisted) and Stage 2 (assisted) for the same reader on the same case constitutes the primary paired observation.

This sequential per-case design mirrors the intended clinical workflow, where a healthcare professional first forms an initial diagnostic impression and then consults the device as a decision support tool. Any carry-over effect from the unassisted diagnosis to the assisted diagnosis is conservative: readers anchored to their Stage 1 diagnosis are less likely to change it in Stage 2, thereby underestimating rather than overestimating the device's incremental benefit. This design has been validated in three prior MRMC studies (SAN_2024, BI_2024, PH_2024), all of which demonstrated statistically significant diagnostic accuracy improvements using the same methodology.

Reference standard (ground truth)

The reference standard for diagnostic accuracy is the published atlas diagnosis — i.e., the diagnosis assigned to each clinical image by the originating public dermatological atlas from which the image was sourced. Each image in the dataset was sourced from a peer-reviewed or institutional dermatological atlas where cases are labelled by expert dermatologists, typically based on clinical-pathological correlation.

The ground truth diagnosis is encoded as an ICD-11 code for each case and was established prior to and independently of this study. The reference standard is not modified or influenced by the device output or by the participating readers' assessments.

Limitations of the reference standard: Atlas diagnoses may not all be histopathologically confirmed. However, dermatological atlas labels represent expert consensus diagnoses and are a widely accepted reference standard in multi-reader multi-case studies evaluating diagnostic decision support tools. The impact of this limitation on accuracy calculations is mitigated by: (1) the large number of cases (149), which reduces the influence of any individual misclassified case; (2) the self-controlled design, where both unassisted and assisted conditions are evaluated against the same reference standard, so any reference standard error affects both arms equally; and (3) the consistency of this approach with the three source studies (SAN_2024, BI_2024, PH_2024), which used the same reference standard methodology.

Population

In this study, the population will consist of dermatologists. A minimum of 5 dermatologists will participate in the study. Each participant will be presented with 149 images to review.

Sample size

This study aims to evaluate whether the use of the device improves diagnostic accuracy by at least 10 percentage points among dermatologists diagnosing conditions on Fitzpatrick phototype V–VI skin.

Power calculation

The sample size was determined using a two-sided McNemar's test for paired binary outcomes (correct/incorrect diagnosis, with and without the device), with the following parameters:

• Baseline accuracy (unassisted): 60%. This is a conservative estimate based on published literature on dermatologist diagnostic accuracy for multi-condition image-based assessments, and is consistent with the unassisted accuracy observed in the source studies (SAN_2024, BI_2024, PH_2024).
• Expected accuracy (assisted): 70%. This represents the minimum clinically meaningful improvement of 10 percentage points.
• Significance level (alpha): 0.05, two-sided.
• Statistical power (1 - beta): 0.80.
• Discordant proportion: Under the assumption that 25% of paired observations are discordant (i.e., one arm is correct and the other is not), a minimum of approximately 200 paired observations is required per McNemar's test (Lachin, 1992).

With 5 readers × 149 images = 745 paired observations, the study is well-powered even after accounting for the within-reader correlation inherent in the MRMC design. Adjusting for intra-reader correlation (ICC ≈ 0.15, based on the source studies), the effective sample size is approximately 460 independent observations, which exceeds the minimum of 200 required.

For the MRMC-specific analysis (Obuchowski-Rockette method), the minimum of 5 readers and 149 cases exceeds the recommendations of Hillis (2011) for detecting reader-averaged differences in diagnostic accuracy of 10 percentage points, given the number of cases available per condition category.

Justification for the image set

The 149 images were selected from 158 unique clinical cases across three completed MRMC studies. Images representing 93 cases from BI_2024, 29 cases from PH_2024, and 27 cases from SAN_2024 were included. Nine cases were excluded during the Gemini conversion process because the safety filter blocked the generation of converted images for both Fitzpatrick phototypes. All included images represent Fitzpatrick phototype V–VI presentations and passed a visual quality control review to ensure clinical fidelity, preservation of lesion morphology, and realistic skin tone presentation. Four additional cases were excluded during visual QC because both Fitzpatrick V and VI conversions failed quality criteria.

Selection of 5 dermatologists

The recruitment of a minimum of 5 dermatologists aims to ensure adequate inter-observer variability while maintaining a manageable and representative cohort. Five readers represent the minimum required for valid MRMC statistical analysis per Obuchowski (2004) and Hillis (2011).

Dermatological conditions and case distribution

The 149 images represent 28 distinct dermatological conditions, including 10 cases of malignant conditions. The distribution is as follows:

ICD-11 Code	Condition	Images	Risk
EA89	Generalised eczematous dermatitis	13	Benign
ED92.0	Hidradenitis suppurativa	11	Benign
ED80.3	Nodular acne	10	Benign
1B72	Impetigo	9	Benign
EA90.40	Generalised pustular psoriasis	8	Benign
1F28	Dermatophytosis	8	Benign
EA90.0	Plaque psoriasis	7	Benign
ED80.Z	Acne, unspecified	7	Benign
2F20.Z	Melanocytic naevus, unspecified	7	Benign
2C30	Melanoma of skin	7	Malignant
2F21.0	Seborrhoeic keratosis	6	Benign
EA81	Seborrhoeic dermatitis	6	Benign
EB05	Urticaria	6	Benign
—	Subcorneal pustular dermatosis	5	Benign
EA90.42	Palmoplantar pustulosis	5	Benign
ED80.4	Severe inflammatory acne	5	Benign
EB40	Pemphigus	5	Benign
—	Acute generalized exanthematous pustulosis	4	Benign
2C32	Basal cell carcinoma of skin	3	Malignant
EA90	Psoriasis	3	Benign
2F20.1	Atypical melanocytic naevus	2	Benign
EK90.0&XH36H6	Actinic keratosis	2	Benign
EA90.4	Pustular psoriasis	2	Benign
ED70	Alopecia	2	Benign
1E91	Herpes zoster	2	Benign
EE12.1	Onychomycosis	2	Benign
EE80.0	Granuloma annulare	1	Benign
EH90.Z	Pressure ulcer	1	Benign

Malignant cases: 10 (6.7% of total). Benign cases: 139 (93.3%).

Duration

The total duration of the study is estimated at 4 months, including the time required after the collection of the corresponding images for the closing and editing of the database, the analysis of the data, and the preparation of the final report of the study.

Acceptance criteria

Note: No acceptance criteria found for study code "MAN_2025" in performanceClaims.ts

Inclusion criteria

• Board-certified dermatologists, regardless of their professional experience.
• High-quality anonymized images of skin conditions representing Fitzpatrick phototype V–VI presentations.

Exclusion criteria

• Low-quality images which cannot be properly analyzed.
• Images where the skin condition morphology has been altered beyond clinical recognition.

Variables

Main variable

The main variable of this study is the diagnostic performance (accuracy, sensitivity, and specificity) in different skin pathologies using the device, measured against the atlas-labelled reference standard diagnosis (ICD-11 code). Performance will be calculated both with and without the use of the device.

• Top-1 diagnostic accuracy: Proportion of cases where the reader's primary diagnosis matches the reference standard.
• Top-3 diagnostic accuracy: Proportion of cases where the reference standard diagnosis appears among the reader's top 3 differential diagnoses.
• Malignancy detection AUC: Area Under the Receiver Operating Characteristic Curve for the binary classification of malignant vs. benign conditions.

Secondary variables

• Percentage of cases that should be referred according to the HCP with the information provided by the device.

Condition of interest

Dermatological conditions presented on Fitzpatrick phototype V–VI skin, as depicted in anonymized clinical images sourced from public dermatological atlases. Twenty-eight distinct conditions are represented, including both malignant (melanoma, basal cell carcinoma) and benign presentations.

Limitations of clinical research

The main limitations of this study include several factors that may influence the perception and effectiveness of the AI-based device. Firstly, the acceptance and trust of healthcare professionals in these emerging technologies can vary significantly. The device's effectiveness may be compromised if users are not fully convinced of its accuracy or usefulness, thereby affecting the overall perception of its performance.

Additionally, image quality is crucial for the device's performance. Issues such as low-quality photographs, errors in cropping lesions, or variations in lighting and focus can deteriorate the quality of the data received by the system, which may negatively influence the evaluation and perception of its effectiveness by the researchers.

Variability in image conditions is also an important aspect to consider. Differences in lighting, colour, shape, size, and focus of the images, along with the number of images available for each patient, can affect the accuracy of the results. High variability in images of the same patient or an insufficient number of representative images can lead to a decrease in the expected diagnostic accuracy of the device.

Furthermore, the consistency of researchers in using the device is crucial. Variations in how diligently investigators use the device can impact the study's findings. If the investigators are not consistent in their use of the device, it can lead to unreliable results and affect the overall assessment of its efficacy.

Another limitation is the Hawthorne effect, where study participants may change their behaviour simply because they know they are being observed. This awareness can influence their decisions and actions within the study, potentially skewing the results and not accurately reflecting how the device would be used in a non-study environment.

Finally, this study is limited to Fitzpatrick phototype V–VI skin presentations. While this targeted approach addresses a specific evidence gap, the results should be interpreted in the context of the broader evidence base that includes phototypes I–IV from prior studies.

Ethical considerations

This study does not require Ethics Committee approval. No patients are involved — all images are fully anonymized and sourced from public dermatological atlases containing no identifiable patient information. The participating healthcare professionals are not research subjects; they are professional evaluators providing diagnostic opinions in a controlled setting. The study is observational and non-interventional: no clinical decisions, treatments, or patient care pathways are influenced by this study.

This exemption is consistent with the regulatory classification described in "Regulatory classification of this study" and with the approach accepted for the three source studies (SAN_2024, BI_2024, PH_2024), which used the same study design and were conducted without Ethics Committee approval under the same justification.

Data confidentiality

All study data are collected and processed in compliance with Regulation (EU) 2016/679 (GDPR) and Spanish Organic Law 3/2018. The study images are fully anonymized and contain no patient-identifiable information. Each participating practitioner is assigned a unique study identification code; the list linking identification codes to practitioner identities is maintained by the Principal Investigator in a secure area with restricted access. No personal data are included in the analysis dataset or in any study report.

Bias minimization measures

In clinical research, minimizing bias is essential to ensure the validity and reliability of the study's results. The following measures are implemented:

• Self-controlled design: Each reader serves as their own control for every case, eliminating between-reader variability as a confounding factor. The paired comparison (unassisted vs. assisted) for the same reader on the same case isolates the device's incremental contribution.
• Progressive disclosure: The sequential presentation of device information within each case ensures that the unassisted diagnosis (Stage 1) is recorded before the reader sees any device output, preventing information leakage from the assisted condition to the unassisted condition.
• Conservative carry-over: Any recall of the Stage 1 diagnosis during Stage 2 biases conservatively, as readers anchored to their initial diagnosis are less likely to adopt the device's suggestion, underestimating rather than overestimating the device's benefit.
• Randomised case order: The order of case presentation is independently randomised for each reader, preventing order effects and ensuring that fatigue or learning effects do not systematically bias results for specific cases.
• Standardized protocol: All participants follow the same procedures for conducting the study and recording outcomes, reducing variability due to differences in how the device is used.
• Prospective data collection: Collecting data prospectively reduces the chance that participants or investigators will inaccurately recall past events.
• Pre-defined endpoints: Before the study begins, primary and secondary outcomes are defined, which prevents selective reporting of only favourable outcomes.

Calendar

The total duration of the study is estimated at 4 months, including the time required after the recruitment of the participating HCPs and collection of all images, and for closing and editing the database, data analysis, and preparation of the final study report.

Monitoring plan

The sponsor will hold a meeting with the investigative team at the beginning of the study to address any potential questions and ensure that data is being collected properly.

The clinical investigation will be monitored by a designated clinical monitor, who is independent of the investigational site and appointed by the sponsor. The purpose of monitoring is to ensure that:

• The rights, safety, and well-being of the subjects are protected.
• The data reported are accurate, complete, and verifiable from source documents.
• The clinical investigation is conducted in compliance with the Clinical Investigation Plan (CIP), applicable regulatory requirements (including Regulation (EU) 2017/745), and ISO14155:2020.

In this way, monitoring will be performed through:

• Remote monitoring activities: Including scheduled video or telephone meetings every 3 months with the investigators to review study progress, discuss challenges, and ensure ongoing compliance.
• On-site visits: If deemed necessary based on data review, protocol deviations, or issues raised during remote monitoring. In this study, this will be carried out online.
• Source data verification (SDV): The sponsor in this case has secure access to anonymised source documents (e.g., image files, clinical assessment records) via encrypted digital platforms or controlled site systems to verify data accuracy and consistency with Case Report Forms (CRFs).
• Risk-based monitoring approach: The extent and frequency of monitoring activities will be adapted to the complexity and risk level of the investigation. Given that this study does not involve significant deviation from routine care, the monitoring will be primarily remote unless specific issues arise.

All monitoring activities will be documented in monitoring visit reports. Any protocol deviations or non-compliance identified will be documented, assessed for impact, and communicated promptly to the sponsor. Corrective actions will be tracked and followed up by the monitor or the sponsor.

Completion of the investigation

After the final closure of the study, a Clinical Investigation Report (CIR: T-015-006 Clinical Investigation Report) will be drafted, even in the event of early termination or suspension. The results obtained (whether positive, inconclusive, or negative) will be documented.

Additionally, if deemed appropriate, the results may be published in scientific journals. All the investigators who approved this clinical investigation will be acknowledged, and any funds received by the author for the study and its source of funding will be disclosed. The anonymity of participants in the clinical investigation will be maintained at all times.

Upon completion of the study, the results of the clinical utility and satisfaction surveys, as per the annexed models, may be presented at conferences and scientific meetings, subject to prior authorization by both parties. Press releases and other communications may also be issued to share the study's results. All publications and communications must be reviewed and approved by the parties involved.

Statistical analysis

Primary analysis (MRMC methodology)

The primary analysis will use the Obuchowski-Rockette (OR) method for multi-reader multi-case studies, which accounts for the crossed design where every reader evaluates every case. This method treats both readers and cases as random effects and correctly handles the correlation structure inherent in the MRMC design.

The primary endpoint — the difference in reader-averaged diagnostic accuracy between the assisted (Stage 2) and unassisted (Stage 1) conditions — will be estimated with 95% confidence intervals using the OR method. Statistical significance will be assessed at alpha = 0.05 (two-sided). The Hillis-Berbaum correction for small numbers of readers will be applied if the number of readers is <10.

For malignancy detection, the area under the ROC curve (AUC) will be computed for each reader in each condition, and the reader-averaged difference in AUC between conditions will be tested using the OR method.

Secondary analyses

Each variable will be characterized using frequency distributions for qualitative variables and central tendency statistics such as mean and median and variability statistics such as standard deviation (S.D.) or interquartile range for quantitative variables according to their distributional characteristics.

The analysis will focus on calculating diagnostic accuracy, sensitivity, and specificity for dermatologists, both with and without the use of the medical device. For each of these metrics, two types of results will be analyzed:

1. The percentage of variation in the metric attributable to the use of the device.
2. The absolute value of the metric, which will be compared against the state of the art and the values obtained by the practitioners during the study.

Additionally, a comparative analysis will be performed against the results of the three source studies (SAN_2024, BI_2024, PH_2024) to evaluate whether diagnostic performance on Fitzpatrick phototype V–VI skin is consistent with the performance observed on phototypes I–IV.

Within-reader comparisons (assisted vs. unassisted) will be conducted using McNemar's test for binary outcomes. GEE models with exchangeable correlation structure will be fitted to account for within-reader clustering when appropriate. Interobserver concordance analyses will be performed by estimating the kappa coefficient.

Subgroup analyses by condition category (malignant vs. benign) and by individual high-prevalence conditions will be performed as exploratory analyses.

Analyses will be performed using appropriate statistical software. Values of p lower than 0.05 will be considered significant.

Data management

All study data will be collected and managed through a centralized, secure web-based platform that serves as the electronic Case Report Form (eCRF). Each participating dermatologist will receive individual login credentials (username and password) to access the platform. The platform will present the standardized set of 149 images to each practitioner and capture their diagnostic assessments and clinical decisions in response to structured questions.

Data collection

Diagnostic assessments, referral decisions, and remote consultation feasibility evaluations will be electronically captured directly in the platform when practitioners provide their responses. All data entries are time-stamped and automatically stored in a secure central database.

Access control and traceability

Access to the platform is restricted to authorized practitioners through individual user credentials. Complete access logs are maintained for each practitioner, documenting login times, data entry activities, and timestamps of all interactions. This audit trail ensures full traceability of who accessed what data and when.

Data anonymity

Study images are completely anonymized and bear no patient identifiers. Each practitioner receives a unique study identification code that is not linked to their personal identity in the analysis dataset. The Principal Investigator maintains a confidential list linking practitioner identification codes to their identities, stored separately in a secure area with restricted access.

Data export and analysis

Once all data collection is completed and the database is closed and locked by the Principal Investigator, data are exported to CSV format for statistical analysis. The exported dataset contains only anonymized image data and practitioner responses, with no personal identifying information.

Data security

All data transmitted to and stored on the platform are encrypted using industry-standard security protocols. Access to the central database is restricted to authorized personnel only, and the system implements appropriate technical and organizational measures to prevent unauthorized access, alteration, or loss of data.

Current legislation will be complied with in terms of data confidentiality protection (European Regulation 2016/679 of 27 April, on the protection of natural persons with regard to the processing of personal data and the free movement of such data, and Spanish Organic Law 3/2018 of 5 December, on the Protection of Personal Data and guarantee of digital rights).

CIP Modification

Any substantial modification to this CIP will be reviewed and approved by the Principal Investigator and the sponsor before implementation. Modifications will be documented with version number, date, and description of changes. Since this study does not require Ethics Committee or competent authority approval, modified protocols do not require resubmission to these bodies.

CIP Deviations

Any deviations from the CIP will be documented and assessed by the Principal Investigator for their impact on the integrity of the study data and the validity of the results. Deviations will be classified as minor or major and will be reported in the Clinical Investigation Report (CIR).

Start, follow-up and end reports

The start of the study will be notified to the principal investigator and all the participant investigators.

Upon obtaining the study conclusions, a final report (T-015-006 Clinical Investigation Report (CIR)) will be prepared and submitted to the sponsor of the study.

Statements of compliance

The sponsor and the principal investigator commit to conducting this study in accordance with the applicable standards and regulations described in the "Compliance Statement" section. The study will be conducted in accordance with the principles of Good Clinical Practice to the extent applicable to this non-interventional, observational study design.

Informed Consent process

For patients/image subjects

Informed consent from patients whose images might appear in the atlases or public sources is not required for this study because: (1) the study utilizes completely anonymized images from public dermatological atlases and publicly available sources where individuals cannot be identified; (2) there is no active patient recruitment or direct involvement of patients in this study; (3) the images used constitute non-personal data under GDPR as they contain no information allowing identification of data subjects; and (4) the study does not involve any intervention, modification of care, or processing of sensitive personal health data.

For healthcare practitioners (dermatologists)

While formal Informed Consent Forms are not required for practitioners (as this study is observational and non-interventional with respect to their clinical practice), all participating dermatologists will receive comprehensive written and oral information about the study before commencing their participation. The information provided will include:

• Study objectives and design.
• Study procedures, including the two-phase assessment process and expected time commitment.
• The voluntary nature of participation and the right to withdraw at any time without justification.
• Data handling practices, including anonymization of practitioner responses and confidentiality protections.
• Sponsor contact information for questions or concerns.

Each practitioner will sign a participation contract with the sponsor that formalizes their involvement, outlines their obligations, and confirms their understanding and voluntary agreement to participate.

Adverse events, adverse product reactions and product deficiencies

Adverse Events (AE) and Adverse Events to the Product (AEP)

An AE is any unintended medical event, unanticipated illness or injury, or unintended clinical signs (including abnormal laboratory findings) in subjects, users, or other persons, whether or not related to the investigational product and whether intended or unintended.

An AEP is an adverse event related to the use of an investigational medical device.

Given these definitions, potential AEPs or AEs are documented in the product's IFU.

Product deficiencies

Possible inadequacies of a medical device may relate to its identity, quality, durability, reliability, safety, or performance.

Product deficiencies in the investigation will be managed by the sponsor according to non-conforming product control procedures. When appropriate, corrective and/or preventive actions will be taken to protect the safety of subjects, users, and other individuals.

Serious Adverse Events, serious adverse events to the product and serious and unexpected adverse events to the product

According to UNE-EN ISO 14155:2021:

• A Serious Adverse Product Reaction (SAEP) is a SAE that has produced any consequence characteristic of a serious adverse event.
• A Serious Adverse Event (SAE) is an AE that results in any of the following events: death, serious deterioration of the health status of the subject, users or other persons, fetal distress, fetal death, congenital anomaly or birth defect.
• A Serious Unexpected Adverse Event to the Product (SUAEP) is a SAE that, due to its nature, incidence, intensity or consequences, has not been identified in the updated risk assessment.

Taking into account these definitions, there are no SAEP, SAEs or SUAEPs related to the use of the product.

Foreseeable adverse events and adverse events to the product

The foreseeable adverse events and expected adverse reactions to the product, as well as their incidence, mitigation and treatment are documented in the T-013-002 Risk management record of the product under study.

Data Monitoring Committee (DMC)

A Data Monitoring Committee will not be established for this study. This decision is justified because: (1) the study is non-interventional and does not affect patient care or safety; (2) no patients are involved — the participants are healthcare professionals providing diagnostic assessments; (3) there are no safety endpoints to monitor; and (4) the study duration and design do not warrant interim analyses for safety or futility.

Suspension or early termination of clinical research

The sponsor may suspend or terminate the study at any time if: (1) the study objectives are unattainable; (2) significant protocol deviations compromise data integrity; or (3) unforeseen circumstances prevent the study from being conducted in accordance with this CIP. In the event of early termination, the sponsor will notify the Principal Investigator and all participating investigators, document the reasons for termination, and prepare a CIR covering all data collected up to the point of termination.

Signature meaning

The signatures for the approval process of this document can be found in the verified commits at the repository for the QMS. As a reference, the team members who are expected to participate in this document and their roles in the approval process, as defined in Annex I Responsibility Matrix of the GP-001, are:

• Author: Team members involved
• Reviewer: JD-003 Design & Development Manager, JD-004 Quality Manager & PRRC
• Approver: JD-001 General Manager

ㅤ