Terms of Service — ProstPlan AI

1. Definitions

• "Service": The ProstPlan AI platform, including the web application, Edge Functions, data extraction API and all associated documentation.
• "User": Qualified healthcare professional (physician, urologist, oncologist) who accesses the Service.
• "Clinical Data": Health information entered or extracted by the Service, including PSA, Gleason, PI-RADS, biopsy sectors and MRI parameters.
• "Outputs": Calculation results, risk estimates, scores, 3D visualisations and reports generated by the Service.
• "Models": Statistical and AI algorithms used by the Service (PRECE, PLND, Trifecta, MSKCC, Tumor Volume, extraction via Gemini).

2. Nature of the Service

2.1 Clinical Decision Support

ProstPlan AI is a clinical decision support (CDS) tool intended exclusively for qualified healthcare professionals. The Service:

• Calculates probabilities and risk estimates based on models published in peer-reviewed literature
• Presents structured data and visualisations to support surgical planning
• Does NOT constitute medical diagnosis
• Does NOT recommend specific treatments
• Does NOT replace clinical judgement of the healthcare professional

2.2 Auxiliary Tool

Per CFM Resolution No. 2.454/2026 (Brazil), EU AI Act (Regulation EU 2024/1689) and FDA guidance for Clinical Decision Support Software (2022), ProstPlan AI acts as a physician assistant. Final clinical decisions remain the exclusive responsibility of the healthcare professional.

3. Eligibility and Access

3.1 Restriction to Healthcare Professionals

The Service is intended exclusively for qualified healthcare professionals with training in urology, oncology or related fields. By using the Service, the User declares that:

• They are a licensed physician in legal practice
• They possess technical knowledge to interpret the Service Outputs
• They will not use the Service for non-clinical purposes without authorisation

3.2 Account and Authentication

• Access requires authentication via OAuth (Google), email/password or magic link
• The User is responsible for the security of their credentials
• Each account is personal and non-transferable
• Unauthorised access attempts will be logged and may result in blocking

4. Physician Responsibility

4.1 Mandatory Human Verification

The User agrees to:

1. Verify all data automatically extracted from PDF reports before using it for calculations
2. Validate each calculation output (PRECE, PLND, Trifecta, MSKCC) before exporting reports or making clinical decisions
3. Not delegate clinical decision-making to the Service
4. Document any changes to data suggested by AI in the clinical record system

4.2 Subsidiary Responsibility

Per CFM Resolution No. 2.454/2026:

• The physician is the primary responsible for any clinical decision based on the Service Outputs
• The Service provider is responsible for fidelity of implementation of published models, data security and transparency about limitations
• In case of discrepancy between Outputs and physician clinical judgement, clinical judgement always prevails

5. Model Limitations

5.1 Statistical Models

5.2 LLM Extraction (Gemini)

• Automatic PDF data extraction uses a Large Language Model (Gemini 2.0 Flash)
• There is a possibility of extraction error (hallucination, non-standard formatting, regional nomenclature)
• Human verification after extraction is mandatory and must not be ignored
• The dual verification system (2 passes) mitigates but does not eliminate the risk of error

6. Intellectual Property

The statistical models implemented in the Service are based on coefficients published in peer-reviewed scientific literature. Rights to the original models belong to their respective authors and institutions. The software implementation (code, interface, 3D visualisations) is the property of the ProstPlan AI provider.

7. Data Protection

Use of the Service is subject to the Privacy Policy, which details data categories, legal basis, security measures, data subject rights and international transfers.

8. Disclaimer

The Service is provided "as is". The provider does not guarantee the accuracy, completeness or suitability of the Outputs for any specific clinical case. To the maximum extent permitted by applicable law, the provider shall not be liable for clinical decisions made based on the Outputs.

9. Changes to the Terms

The Terms may be updated periodically. Material changes will be communicated via in-app modal and/or email. Continued use of the Service after notification constitutes acceptance of the new Terms.

10. General Provisions

These Terms are governed by applicable legislation according to the User's jurisdiction (Brazil: LGPD, CFM; EU: GDPR, AI Act; USA: HIPAA, FDA CDS). For questions about these Terms, contact: [insert contact]

Version History