Cutaneous Lupus Erythematosus Severity Endpoints for Clinical Trials

The AI scoring provided by Legit.Health automates the visual components of the CLASI (Cutaneous Lupus Erythematosus Disease Area and Severity Index), the reference outcome measure in cutaneous lupus erythematosus (CLE) drug development. It quantifies the erythema, scale, induration, pigment loss and hair loss that CLASI is built from, from standard smartphone images, and assembles them into the CLASI Activity and Damage scores.

Example per-visit CLASI report. Synthetic imagery, real device outputs for the signs shown; the full report set is on the Sample outputs page.

The CLASI instrument

CLASI is the validated reference instrument for CLE, developed by Albrecht and colleagues (2005) and adopted as the outcome measure in cutaneous lupus trials. It is unusual among dermatological indices because it produces two independent scores from one examination:

CLASI-A

Activity

Range0–70

Erythema and scale/hypertrophy per region, plus alopecia and mucosal involvement

BehaviourReversible

Responds to treatment within weeks; the usual efficacy endpoint

CLASI-D

Damage

Range0–56

Dyspigmentation and scarring/atrophy per region, plus scalp scarring

BehaviourCumulative

Accumulates irreversibly; the long-term burden measure

Both scores come from the same set of photographs, and separating activity from damage reliably is what makes each one interpretable. The instrument's own validation and the literature behind it are on the Clinical evidence and Publications pages.

Endpoint capabilities

The platform provides four configurable endpoints, tailored to the study protocol:

Endpoint	Definition	AI output	Typical use in protocol
CLASI-A (Activity)	Reversible disease activity: erythema and scale/hypertrophy per region, plus alopecia and mucosal involvement	Activity score 0–70, from measured signs plus clinician-entered items	Primary
CLASI-D (Damage)	Irreversible damage: dyspigmentation and scarring per region, plus scalp scarring	Damage score 0–56	Secondary
Depigmentation extent	Continuous depigmented area per lesion and region; the basis of a repigmentation measure the binary CLASI-D item cannot express	Affected area, relative or mm² with marker capture	Exploratory
Per-sign intensity and extent	Erythema, scale and induration as continuous per-region measures, at higher resolution than the CLASI item scales	Per-sign value and area per region	Exploratory

Why the continuous measures matter

CLASI is a coarse instrument by design: erythema is a four-point item, dyspigmentation a single present-or-absent flag. The device measures the same signs continuously, so a study can report the CLASI scores regulators expect and the higher-resolution per-sign measures that detect a treatment effect earlier. The clearest case is repigmentation: a depigmented plaque refilling with pigment is a change patients value, and one CLASI-D structurally cannot show, but a continuous depigmentation extent can.

What the AI measures

Each measurement is produced from one photograph of one anatomical region.

Measurement	Output	CLASI component it serves
Erythema intensity	Graded severity of redness	CLASI-A erythema
Erythema extent	Affected area, relative or absolute	Lesion burden within the region
Desquamation intensity	Graded severity of scaling	CLASI-A scale
Induration intensity	Graded plaque thickening	CLASI-A hypertrophy
Depigmentation extent	Affected area, relative or absolute	CLASI-D dyspigmentation
Lesion area	Physical area in mm², with marker capture	Longitudinal lesion measurement
Hair loss percentage	Affected proportion of the scalp	CLASI-A and CLASI-D scalp items
Diagnosis support	Ranked differential including CLE	Screening and eligibility review
DIQA image quality	Quality score with accept or recapture	Applies to every capture
Skin and body segmentation	Pixel-level masks	Underlies every extent figure

The full item-by-item mapping, including the CLASI components that are not derived from an image, is on the Scoring methodology page.

How the AI works

The pipeline runs in three stages, from photograph to assembled score.

Stage 1: Image quality gate

Every image is assessed by DIQA (Dermatology Image Quality Assessment) before any sign is read. Captures whose exposure, focus or framing fall outside the acceptable range are rejected for recapture, so a poor photograph produces a prompt rather than a wrong number. See the Imaging protocol.

Stage 2: Per-sign measurement

Dedicated models measure each visual sign in the photographed region: erythema, desquamation and induration for activity; depigmentation for damage; hair loss on the scalp. Each sign is returned with a confidence figure and a pixel-level segmentation mask.

An extent is reported two ways. Relative extent divides the segmented sign area by the photographed region:

E_{\text{rel}} = \frac{A_{\text{sign}}}{A_{\text{region}}}

Where a calibration marker of known side $d$ (mm) spanning $p$ pixels is in frame, the segmented pixel count $N_{\text{px}}$ converts to absolute area in physical units:

A_{\text{mm}^2} = N_{\text{px}} \left(\frac{d}{p}\right)^2

The absolute form is independent of working distance, which is what makes it comparable across visits; the imaging protocol covers marker capture. Every sign is also qualified by the scoring system that produced it, because the same sign is scored on different scales by different validated instruments and the values are not interchangeable; see scale alignment.

Stage 3: CLASI assembly

Per-region sign values are summed into the two CLASI scores. The activity score adds the two activity signs across the scored anatomical regions, plus the three clinician-entered items:

\text{CLASI-A} = \sum_{\text{regions}} (E_r + S_r) + M + H + A_{\text{np}}

with erythema $E_r$ (0–3), scale or hypertrophy $S_r$ (0–2), mucous-membrane involvement $M$ (0–1), recent hair loss $H$ (0–1) and non-scarring alopecia $A_{\text{np}}$ (0–3); total range 0–70.

The damage score adds dyspigmentation and scarring, with the dyspigmentation subtotal doubled when it persists beyond twelve months:

\text{CLASI-D} = (1 + p)\sum_{\text{regions}} D_r + \sum_{\text{regions}} C_r + C_{\text{scalp}}

with dyspigmentation $D_r$ (0–1), the persistence flag $p \in \{0, 1\}$ , scarring or atrophy $C_r$ (0–2) and scalp scarring $C_{\text{scalp}} \in \{0, 3, 4, 5, 6\}$ ; total range 0–56.

The device measures the imaged terms ( $E_r$ , $S_r$ , $D_r$ , and the alopecia extent behind $A_{\text{np}}$ ); the investigator supplies $M$ , $H$ and the persistence flag $p$ at review. Assembly depends on each image being attached to a patient, a visit and a named region; without that, the device returns valid per-image values that cannot be summed. The Imaging protocol sets out the metadata this requires.

CLASI-A severity strata

Activity scores map to severity bands validated by Klein and colleagues (2011):

CLASI-A score	Severity
0–9	Mild
10–20	Moderate
21–70	Severe

The same work found that a clinically meaningful response corresponds to roughly a 4-point or 20% fall in the activity score, the basis for a CLASI-A response endpoint. Details and references are on the Clinical evidence page.

Protocol flexibility

The region set, visit structure, target lesions and calibration are configured per study at setup. A study can capture all 13 regions at every visit or take a comprehensive baseline with focused follow-up, and can report absolute lesion area in mm² where calibration markers are used. The Imaging protocol and Trial workflow pages cover the options.

📄️Scoring Methodology

Item-by-item technical detail on how the AI quantifies the CLASI Activity and CLASI Damage components, the 13-region structure, scale alignment between sign models and CLASI item scales, and how per-image values aggregate into a CLASI score.

📄️Imaging Protocol

Image capture protocol for cutaneous lupus erythematosus clinical trials, covering the 13-region CLASI capture set, target lesions, the metadata required to assemble a CLASI score, DIQA quality control, colour and scale calibration for erythema, pigmentation and lesion area, and facial anonymization.

📄️Clinical Evidence

The evidence base for automated CLASI in cutaneous lupus erythematosus. The CLASI instrument's own validation, the component models' validation in their source indications, feasibility findings on CLE imagery, and the validation pathway.

📄️Trial Workflow

End-to-end workflow for cutaneous lupus erythematosus severity endpoints, covering retrospective image analysis versus prospective CLASI reconstruction, protocol design, capture, scoring, review, longitudinal tracking, and EDC data export.

📄️Sample Outputs

What sponsors, CROs and investigators receive from a cutaneous lupus erythematosus assessment. A per-visit report with real device outputs, the repigmentation signal CLASI-D cannot express, segmentation overlays, and structured EDC export fields.

📄️Publications

Key references for automated CLASI in cutaneous lupus erythematosus. The CLASI instrument's development, validation and responsiveness literature, and the peer-reviewed methods behind the device's component measurements.

📄️Limitations

Transparent documentation of what AI CLASI component scoring can and cannot do in cutaneous lupus erythematosus, covering the items not derived from an image, the scarring and alopecia boundaries, region traceability, skin phototype, and CLE-specific validation status.

The CLASI instrument​

Endpoint capabilities​

What the AI measures​

How the AI works​

Stage 1: Image quality gate​

Stage 2: Per-sign measurement​

Stage 3: CLASI assembly​

CLASI-A severity strata​

Protocol flexibility​

Further reading​