Psoriasis Severity Endpoints for Clinical Trials

The AI scoring provided by Legit.Health delivers automated, clinically validated psoriasis severity scoring for clinical trials, quantifying erythema, desquamation, induration, and affected body surface area across four body regions to compute the Psoriasis Area and Severity Index (PASI).

Psoriasis Area and Severity Index

APASI Score: 13.6

Report Information

☐

Timestamp

3/24/2026, 2:17:03 AM

☐

Analysis performed in

1.8 seconds

☐

Status

Not reviewed

Body region	Weight	Erythema	Desquamation	Induration	BSA %	Area score	Region PASI
Head	0.1	2	2	1	30%	2	1.0
Trunk	0.3	3	2	2	15%	2	4.2
Upper extremities	0.2	2	1	2	10%	1	2.0
Lower extremities	0.4	3	3	2	20%	2	6.4

PASI = 1.0 + 4.2 + 2.0 + 6.4 = 13.6

Regulatory context

The FDA’s guidance document Psoriasis: Developing Drug Products for Treatment. Guidance for Industry. (2022) establishes the standard for clinical trial endpoints:

• PASI is the most widely used clinical tool for measuring psoriasis severity and treatment response in clinical trials.
• The primary efficacy endpoint should be based on the proportion of subjects achieving PASI 75 (75% improvement from baseline) or higher thresholds (PASI 90, PASI 100).
• IGA (Investigator Global Assessment) should be used as a co-primary endpoint, typically measuring the proportion achieving IGA 0 (clear) or 1 (almost clear).
• The assessment should include evaluation of erythema, induration, and desquamation across defined body regions.

PASI is the most widely used severity measure in psoriasis clinical trials worldwide. PASI response rates (PASI 75, PASI 90, PASI 100) are the standard primary efficacy endpoints for biologic and systemic therapies.

Endpoint capabilities

Endpoint	Definition	AI output	Typical use in protocol
APASI score	Automated Psoriasis Area and Severity Index — AI-computed PASI combining erythema, desquamation, induration intensity and affected area across 4 body regions	PASI score 0–72	Co-primary
Erythema intensity	Redness severity per body region, measured by AI from clinical images	Ordinal score 0–4 per region	Secondary
Desquamation intensity	Scaling severity per body region, measured by AI from clinical images	Ordinal score 0–4 per region	Secondary
Induration intensity	Plaque thickness per body region, measured by AI from clinical images	Ordinal score 0–4 per region	Secondary
BSA (Body Surface Area)	Percentage of body surface affected by psoriasis, measured by AI segmentation across 4 body regions	Percentage 0–100% per region	Co-primary

What makes APASI different from manual PASI?

Manual PASI requires a dermatologist to visually estimate four components across four body regions, totalling 16 individual assessments per patient. The affected area (BSA) estimation is particularly subjective and is the largest source of inter-rater variability. APASI automates all 16 assessments using AI, with pixel-level BSA measurement replacing subjective estimation.

How the AI works

PASI is a composite score built from four components measured across four body regions. The AI automates each component independently:

The PASI formula

$$\text{PASI} = \sum_{r \in \{h,t,u,l\}} w_r \cdot A_r \cdot (E_r + D_r + I_r)$$

where $w_r$ is the region weight (head 0.1, trunk 0.3, upper extremities 0.2, lower extremities 0.4), $A_r$ is the affected area score, and $E_r$, $D_r$, $I_r$ are the erythema, desquamation, and induration intensity scores.

PASI score	Severity	Clinical description
0	Clear	No psoriasis (PASI 0)
5	Mild	Mild psoriasis (PASI 1–5)
10	Moderate	Moderate psoriasis (PASI 5–10)
20	Severe	Severe psoriasis (PASI 10–20)
72	Very severe	Very severe psoriasis (PASI >20, maximum 72)

AI component models

Each PASI component is measured by a dedicated AI model:

1. Erythema intensity model: Assesses redness of psoriatic plaques on a 0–4 ordinal scale per region. Production RMAE: 0.13 (expert inter-rater: 0.14).
2. Desquamation intensity model: Assesses scaling severity. Production RMAE: 0.153 (expert inter-rater: 0.17).
3. Induration intensity model: Assesses plaque thickness. Production RMAE: 0.151 (expert inter-rater: 0.17).
4. Erythema surface quantification model: Segments psoriatic areas at pixel level to compute affected BSA percentage. Production IoU: 0.61.

All intensity models perform at or below expert inter-rater variability, meaning the AI is at least as reliable as a dermatologist.

Visual report

The AI produces a structured PASI report for each assessment, with per-region breakdowns:

Psoriasis Area and Severity Index

APASI Score: 13.6

Report Information

☐

Timestamp

3/24/2026, 2:17:03 AM

☐

Analysis performed in

1.8 seconds

☐

Status

Not reviewed

Body region	Weight	Erythema	Desquamation	Induration	BSA %	Area score	Region PASI
Head	0.1	2	2	1	30%	2	1.0
Trunk	0.3	3	2	2	15%	2	4.2
Upper extremities	0.2	2	1	2	10%	1	2.0
Lower extremities	0.4	3	3	2	20%	2	6.4

PASI = 1.0 + 4.2 + 2.0 + 6.4 = 13.6

Protocol flexibility

Protocol	Perspectives	Use case
Standard full-body (11-perspective)	11 views: Head front, Head back, Trunk front, Trunk back, Legs front, Legs back, Foot back, Close-up: head, Close-up: trunk, Close-up: arms, Close-up: legs	Full-body BSA assessment for moderate-to-severe plaque psoriasis. 7 full-body perspectives for area measurement + 4 close-ups for intensity scoring. Supports decentralised (home-based) capture.
Reduced (4-region close-ups only)	4 views: Close-up: head, Close-up: trunk, Close-up: arms, Close-up: legs	Close-up images only for severity intensity assessment without BSA measurement. Suitable for studies where BSA is assessed manually.
Custom	Any combination of perspectives	Any combination of perspectives, defined in collaboration with the sponsor during protocol design.

Decentralised capture

The 11-perspective protocol is designed for decentralised (home-based) image capture. Patients can capture their own images using the Legit.Health mobile application with perspective guidance, allowing PASI assessment without requiring an in-clinic visit.

This has been validated in a real-world Phase 3 clinical trial (JNJ-77242113) for moderate to severe plaque psoriasis.

Further reading

Scoring Methodology

Deep technical detail on each AI component model: erythema, desquamation, induration, and BSA segmentation.

Imaging Protocol

11-perspective full-body capture, decentralised imaging, DIQA quality control, and site standardisation.

Clinical Evidence

APASI publication, production model performance, validation metrics, and regulatory pathway.

Trial Workflow

End-to-end workflow from protocol design to EDC integration, including decentralised capture.

Publications

APASI peer-reviewed paper, DIQA validation, and key references from the psoriasis literature.