Zum Hauptinhalt springen

AI-Powered Endpoint Scoring for Dermatology Clinical Trials

Objective, reproducible dermatological endpoint scoring, from smartphone images, in under 2 seconds.

Acne Lesion And Density INdex

Score: 3.5

Report Information

☐

Timestamp

3/24/2026, 2:17:34 AM

☐

Analysis performed in

1.2 seconds

☐

Status

Not reviewed

Analyzed Left diagonal
☐

Body site

Left diagonal

☐
p]:mb-0>

Image quality

92%

☐
p]:mb-0>

Lesion count

36

☐
p]:mb-0>

Density

0.55

☐
p]:mb-0>

Local score

3.42

Analyzed Right diagonal
☐

Body site

Right diagonal

☐
p]:mb-0>

Image quality

88%

☐
p]:mb-0>

Lesion count

35

☐
p]:mb-0>

Density

0.6

☐
p]:mb-0>

Local score

3.49

Legit.Health provides automated scoring of dermatological endpoints for clinical trials. The AI quantifies severity using established clinical scales (IGA for acne, PASI for psoriasis, SCORAD for atopic dermatitis, and more), eliminating inter-rater variability and reducing site burden.

The technology is embedded in a CE-marked Class IIa medical device (MDD 93/42/EEC), built on an ISO 13485-certified quality management system and compliant with 21 CFR Part 11 for electronic records.

Proven in clinical trials​

130+Investigator sites
12+Countries
Phase 2 & 3Trial phases
4–8 weeksTo deployment

The technology is actively deployed in clinical trials with top-10 pharma companies. Active engagements span psoriasis, acne, alopecia, atopic dermatitis, and generalized pustular psoriasis, including fully decentralised (home-based) capture validated in a pivotal Phase 3 programme.

AI-computed endpoints have been accepted by regulators as secondary endpoints and as a means to detect adverse events in clinical submissions.

See our clinical trial experience →

The problem with manual scoring​

✗

Manual scoring

Same patient, same image — three dermatologists, three different scores:

Dermatologist A
IGA 2(Mild)
Dermatologist B
IGA 3(Moderate)
Dermatologist C
IGA 1(Almost clear)

Typical inter-rater agreement: Cohen's κ = 0.41–0.60

✓

AI scoring

Same patient, same image — always the same result:

Site 1, Visit 1
IGA 2(Mild)
Site 2, Visit 1
IGA 2(Mild)
Site 1, re-scored
IGA 2(Mild)

Intra-rater variability: 0.00 — perfectly reproducible

⇅
Inter-rater variability

Different investigators assign different scores to the same patient

⏱
Slow and costly

Lesion counting takes 5\u201315 minutes per patient and is impractical at scale

📊
Inflated sample sizes

Scoring noise masks treatment effects, requiring larger enrolments

🎓
Site training burden

Calibration exercises are costly, time-consuming, and imperfect

Manual scoring
vs
AI scoring
5–15 min
Time per patient
< 2 sec
Significant
Inter-rater variability
Zero
Requires calibration
Multi-site consistency
Inherent
Limited by raters
Scalability
Unlimited
Per-patient fees
Cost model
Fixed per study

What we provide​

Legit.Health delivers three core capabilities for dermatology clinical trials:

CapabilityWhat it measuresOutput
Automated lesion detectionIndividual lesions via object detection, with per-lesion classificationPer-perspective and total integer counts
Spatial analysisLesion clustering, affected body surface area, and regional distributionDensity ratio or BSA percentage
Severity scoringGlobal severity aligned with established clinical scales (e.g., IGA, PASI, SALT)Score on the reference scale

These three capabilities form the foundation for all condition-specific endpoints. The specific outputs are tailored to each indication (lesion counts and IGA for acne, component scores and PASI for psoriasis, BSA and SALT for alopecia), but the underlying architecture is consistent.

How it works​

The clinical trial workflow follows three steps:

1. Image capture at the investigator site​

Investigators capture standardized photographs using the Legit.Health mobile application on a standard smartphone. The application guides the capture process with condition-specific perspective protocols and validates image quality in real time using the DIQA algorithm.

The number and type of perspectives vary by indication (2 facial perspectives for acne, 11 full-body perspectives for psoriasis, 4 scalp quadrants for alopecia), but the capture experience is consistent across all conditions.

2. Automated AI scoring​

Uploaded images are processed by condition-specific deep learning models. Processing takes under 2 seconds. The system outputs:

  • Lesion detection or segmentation with visual overlays
  • Per-region and global severity scores aligned with the reference clinical scale
  • Image quality assessment (DIQA score)

Acne Lesion And Density INdex

Score: 3.5

Report Information

☐

Timestamp

3/24/2026, 2:17:34 AM

☐

Analysis performed in

1.2 seconds

☐

Status

Not reviewed

Analyzed Left diagonal
☐

Body site

Left diagonal

☐
p]:mb-0>

Image quality

92%

☐
p]:mb-0>

Lesion count

36

☐
p]:mb-0>

Density

0.55

☐
p]:mb-0>

Local score

3.42

Analyzed Right diagonal
☐

Body site

Right diagonal

☐
p]:mb-0>

Image quality

88%

☐
p]:mb-0>

Lesion count

35

☐
p]:mb-0>

Density

0.6

☐
p]:mb-0>

Local score

3.49

3. Scored endpoints for data export​

Results are available through the Legit.Health clinical trials platform and can be exported to EDC systems. Each report includes timestamped severity scores, per-perspective analyses, quality metrics, and longitudinal severity evolution.

Severity chart

Why sponsors choose this​

No special hardware​

Standard smartphone cameras replace expensive specialised photography equipment. This is a fraction of the cost of systems like Canfield VISIA, with no per-site rental fees and no hardware logistics.

Eliminates inter-rater variability​

The AI produces the same score for the same image, every time, at every site, without calibration exercises. This reduces noise in endpoint data and can reduce required sample sizes.

Regulatory-aligned endpoints​

The system provides the standard clinical scales that regulators expect (IGA for acne, PASI for psoriasis, SALT for alopecia) with the added benefit of perfect reproducibility. Novel composite scores (e.g., ALADIN for acne, APASI for psoriasis) are positioned for exploratory and secondary use, adding resolution for Phase 2 dose-finding and internal go/no-go decisions.

Regulatory endpoint acceptance​

AI-computed endpoints from Legit.Health have been accepted as secondary endpoints and for adverse event detection in regulatory submissions. For primary registration endpoints, the system provides the standard clinical scales that regulators expect, with the added benefit of perfect reproducibility.

Under 2 seconds​

Real-time scoring at the investigator site. No images sent to a central reader. No delays waiting for scores.

Multi-site consistency​

Scores are comparable across all investigator sites without calibration exercises, training sessions, or inter-rater reliability assessments. The AI is the rater, and it is perfectly consistent.

QuantifiCare partnership​

Seamless integration with QuantifiCare's standardised photography platforms and clinical trial services. The combined offering provides high-quality standardised photography with AI-powered endpoint scoring, a best-of-breed solution for sponsors who require both.

Indications​

→

Acne

Automated lesion counting, spatial density, and IGA scoring. 2-perspective facial imaging with configurable protocols.

→

Psoriasis

APASI scoring with 11-perspective full-body imaging. PASI 0–72 scale. Erythema, desquamation, induration, and BSA across 4 regions.

⏳

Atopic Dermatitis

ASCORAD scoring. SCORAD 0–103 scale. Coming soon.

→

Alopecia Areata

ASALT scoring with 4-quadrant scalp imaging. SALT 0–100 scale. Automated severity assessment and adverse event monitoring.

⏳

Hidradenitis Suppurativa

AIHS4 scoring. IHS4 scale. Coming soon.

⏳

Additional Indications

Vitiligo (VASI), Urticaria (UAS7), Wounds (AWOSI), Melasma (MASI), and more. Coming soon.

Learn more​

Clinical Trial Experience

Phase 2 and Phase 3 deployments across psoriasis, acne, alopecia, and more. 130+ sites, 12+ countries, top-10 pharma.

Compliance & Data Security

ISO 13485, ISO 27001, GDPR, HIPAA, 21 CFR Part 11. Audit trails, data residency, model locking.

Performance Across Skin Types

Validated across Fitzpatrick I–VI. 280,000+ training images with published stratified performance metrics.

Regulatory status​

CE MarkClass IIaMDD 93/42/EEC
21 CFR Part 11CompliantElectronic records
ISO 13485CertifiedQuality management
ISO 27001CompliantInformation security

Software lifecycle per IEC 62304, usability per IEC 62366-1, risk management per ISO 14971, clinical evaluation per MEDDEV 2.7/1 Rev 4 and MDR Annex XIV.

Jetzt starten

KI-gestützte Dermatologie, validiert durch peer-reviewed Forschung. Vertraut von führenden Krankenhäusern in Europa. Füllen Sie das Formular aus, um zu sehen, wie unsere CE-gekennzeichnete Plattform Ihre Praxis transformieren kann.