Pellivis is a student-led research project exploring the use of artificial intelligence to support preliminary skin-lesion risk assessment from smartphone images.

This page explains the current status of the project, its limitations, and the way in which the Pellivis website, model and application prototype should be understood.

1. Research prototype only

Pellivis is currently a research prototype.

It is being developed to explore whether AI-assisted image analysis may help people become more aware of potentially concerning skin lesions and encourage earlier professional medical review.

Pellivis is not a diagnostic tool. It does not diagnose cancer, melanoma, or any other medical condition.

Any output, result, score, risk indication, image assessment or other information provided by Pellivis should be treated as preliminary and informational only.

2. Not a substitute for medical advice

Pellivis is not a substitute for professional medical advice, diagnosis, treatment, screening or clinical examination.

Anyone who is concerned about a mole, skin lesion, skin change, or any other health issue should contact a qualified healthcare professional, such as a GP, dermatologist or other appropriate clinician.

Users should not delay seeking medical advice, ignore medical advice, or change any medical treatment because of information provided by Pellivis.

If a mole or skin lesion is changing, bleeding, painful, itching, growing, irregular in shape or colour, or otherwise concerning, users should seek professional medical advice promptly.

3. Regulatory status

Pellivis is not currently a CE-marked or UKCA-marked medical device.

Pellivis is not currently approved, certified or authorised for clinical use, diagnosis, treatment, screening, triage or medical decision-making.

The project is not currently available for use as a regulated medical device and should not be used in any clinical pathway unless and until appropriate regulatory, clinical, legal and governance requirements have been satisfied.

4. Current project status

The current status of Pellivis is as follows:

Pellivis is being developed as an early-stage research project and remains subject to further technical testing, clinical evaluation, validation, regulatory review and governance assessment.

5. Limitations

Pellivis has important limitations.

The model has not been validated for diagnosis. It may produce incorrect, incomplete or misleading results.

The accuracy and usefulness of any AI-generated output may be affected by many factors, including:

• image quality;
• lighting conditions;
• camera focus;
• distance from the skin;
• skin tone;
• lesion type;
• lesion location;
• image artefacts;
• similarity between different skin conditions;
• limitations in the training data;
• limitations in the model architecture;
• differences between public datasets and real-world use.

Pellivis may not perform equally across all skin tones, age groups, lesion types, image sources or real-world settings.

False negatives and false positives are possible.

A false negative may occur if the system indicates low risk when a lesion may in fact require medical review.

A false positive may occur if the system indicates higher risk when a lesion is benign or not clinically concerning.

Because of these limitations, Pellivis must not be relied upon to decide whether a skin lesion is cancerous, benign, safe, urgent or non-urgent.

6. No clinical decision-making

Pellivis should not be used by patients, clinicians, healthcare organisations or any other person as the sole basis for clinical decision-making.

This includes decisions about:

• whether to seek medical advice;
• whether to refer a patient;
• whether to delay a referral;
• whether to reassure a patient;
• whether to diagnose or exclude cancer;
• whether to prioritise or deprioritise care;
• whether to begin, stop or change treatment.

Any future clinical use would require appropriate validation, risk management, regulatory assessment, clinical governance, data protection compliance and professional oversight.

7. Data and privacy

Any future collection, storage, review or use of skin images or related health information must be handled with appropriate consent, privacy safeguards and data governance.

Skin images and health-related information may be sensitive personal data. Pellivis should only be used in research, testing or development settings where appropriate permissions, safeguards and oversight are in place.

8. Collaboration and feedback

Pellivis is seeking responsible support from people and organisations with relevant expertise.

In particular, the project is seeking:

• clinical feedback;
• dermatology input;
• regulatory guidance;
• medical-device development advice;
• responsible AI guidance;
• research collaborators;
• data governance advice;
• external validation support;
• responsible pilot-design advice.

The purpose of seeking this support is to ensure that Pellivis is developed carefully, safely and responsibly.

9. No guarantee of accuracy or outcome

Pellivis makes no guarantee that its model, website, application prototype or any related material is accurate, complete, reliable, clinically valid or suitable for any particular purpose.

No guarantee is made that Pellivis will detect, identify, classify, predict or rule out skin cancer or any other condition.

No guarantee is made that use of Pellivis will improve health outcomes, reduce waiting times, prevent disease, prevent death or result in earlier diagnosis.