PASI
Psoriasis Area and Severity Index
Validated on plaque psoriasis presentations across Fitzpatrick 3–6. Automated area and severity component scoring with erythema calibration for darker skin tones.
AICC > 0.85
For Pharma & Clinical Research
AI-based dermatology scoring. Validated on Bhartiya skin.
PASI, EASI, and IGA scoring with population-specific validation on Fitzpatrick 3–6 skin tones. CDSCO-compatible. 21 CFR Part 11 audit trail.
ICC > 0.85PASI concordance
Fitzpatrick 3–6Validation range
21 CFR Part 11Audit trail
The Problem
The Bharatiya skin validation gap.
79% of dermatology datasets originate from populations with Fitzpatrick 1–2 skin tones. Erythema detection, lesion boundary delineation, and severity grading all behave differently on Fitzpatrick 3–6 skin. A model validated only on light skin cannot serve as a clinical endpoint in a Bharatiya trial population. Population-specific validation is not optional — it is a regulatory expectation.
79% of training data = FST 1–2
Western Skin
Under represented in training data
Bhartiya Skins
Regulatory context
The FDA, EMA, and CDSCO increasingly expect AI/ML-based clinical tools to demonstrate validation on the target population. For Bharatiya dermatology trials, this means validation on Fitzpatrick 3–6 skin tones with documentation that satisfies regulatory review. Rupam.ai provides this population-specific validation as a standard part of the clinical package.
Capabilities
Validated scoring indices.
Built on clinically validated scoring indices to ensure accurate and consistent skin assessments. Deliver standardized results you can trust for better diagnosis, tracking, and treatment decisions.
IGA
Investigator Global Assessment
5-point scale scoring validated against dermatologist consensus. Trained to account for post-inflammatory hyperpigmentation common in Bharatiya skin.
ICC > 0.82
EASI
Eczema Area and Severity Index
Body region-specific scoring with separate validation for atopic dermatitis presentations on Bharatiya skin. Accounts for lichenification patterns prevalent in South Asian populations.
ICC > 0.80
Additional indices including mSWAT (mycosis fungoides), BSA, and custom composite endpoints are available on request. Contact our clinical team for protocol-specific scoring requirements.
Integration
Clinical trial integration.
Integrate seamlessly with clinical trial workflows to capture standardized data and track patient outcomes over time. Enable more accurate analysis, better compliance, and faster, evidence-backed insights.
- Step-1
Site onboarding
Standardised imaging protocol distributed to all trial sites. Device-agnostic capture with calibration validation.
- Step-2
Patient imaging
Site staff capture images using the Rupam.ai clinical imaging app. Automatic quality checks ensure usable input.
- Step-3
Central AI scoring
Images are scored centrally by the AI model. Blinded assessment workflow ensures investigator separation from AI output.
- Step-4
Data export
Scores exported in CDISC-compatible format. Full audit trail included. Direct integration with EDC systems available.
Regulatory-ready documentation package.
Comes with a comprehensive documentation package aligned with regulatory requirements and audit standards. Simplify approvals, ensure compliance, and stay ready for inspections without extra effort.
Validation report with ICC scores and methodology
Model version control and change documentation
21 CFR Part 11 compliant audit trail
DPDPA compliance documentation
Data Processing Agreement (DPA) template
NDA-gated technical dossier for regulatory review
The Bhartiya dermatology pharma market.
Bharat is the world's largest dermatology market by patient volume. With over 200 million people affected by skin conditions and a growing pharmaceutical pipeline focused on psoriasis, atopic dermatitis, and vitiligo, the demand for validated clinical endpoints has never been higher.
CDSCO has approved a growing number of biologics and biosimilars for dermatological indications in recent years, each requiring robust efficacy endpoints. As global pharma companies expand Bharatiya trial sites, the expectation of AI-assisted clinical scoring validated on local populations is becoming standard practice.
Sponsors conducting multi-site trials across Bharat need centralised, reproducible scoring that eliminates inter-rater variability while meeting the evidentiary standards of both CDSCO and international regulatory bodies. This is exactly what Rupam.ai's clinical scoring platform provides.
Western Ai modelRupam.Ai model
Built for Bharatiya skin. From the ground up.
Every model, every validation dataset, every scoring algorithm has been designed and tested specifically for Fitzpatrick 3–6 skin tones. This is not a Western model with a regional patch. It is purpose-built clinical AI for the Bharatiya population.
Frequently asked questions
CDSCO does not currently mandate or prohibit AI-generated scoring. However, AI-assisted scoring is increasingly accepted as a supportive tool when accompanied by a validation report, clear methodology documentation, and audit trail. We provide all three. Our validation documentation is designed to satisfy the evidentiary requirements of CDSCO, FDA, and EMA regulatory submissions.
Request our validation documentation.
Our technical dossier includes validation methodology, ICC scores, model architecture, and regulatory compliance documentation. Available under NDA for qualified pharma and CRO teams.