Clinical Trial Workflow and Data Integration
This page describes the end-to-end workflow for cutaneous lupus erythematosus severity assessment using the Legit.Health platform, from protocol design through to EDC data delivery.
Two study shapes
CLE engagements arrive in two forms, and they deliver different things. Establishing which one applies is the first conversation, because it determines whether a CLASI endpoint is available at all.
| Retrospective image analysis | Prospective CLASI study | |
|---|---|---|
| Input | An existing image set, often captured outside a scoring protocol | Capture designed around the 13 CLASI regions |
| Metadata | Frequently no region label, sometimes no patient or visit link | Patient, visit and region attached at capture |
| Delivers | Per-image sign measurements: erythema intensity and extent, desquamation, induration, depigmentation extent, hair loss percentage | The above, plus assembled CLASI-A and CLASI-D per visit |
| Longitudinal analysis | Only where patient and visit links survive | Full trajectory from baseline |
| Typical use | Feasibility, model behaviour on the sponsor's own images, evidence for a prospective design | Endpoint generation |
A retrospective analysis is a genuine and often sensible first step: it demonstrates model behaviour on the sponsor's own population and imaging conditions, at a fraction of the cost and time of a prospective study, and it produces the evidence needed to design one. What it cannot do is produce a CLASI score from images that were never labelled by region. That constraint is a property of the instrument, not of the technology, and it is worth stating at the outset rather than discovering at analysis.
The remainder of this page describes the prospective workflow.
Workflow overview
- Protocol design: configure the study with Legit.Health
- Site setup: deploy the platform to investigator sites
- Patient enrollment: register patients under pseudonymised identifiers
- Image capture: guided region-by-region photography at the site
- AI scoring: automated measurement of the visual CLASI components
- Report review: investigator reviews and completes the assessment
- Longitudinal tracking: CLASI-A and CLASI-D trajectories with configurable alerts
- Data export: structured delivery to the sponsor's EDC
1. Protocol design
| Configuration | Options |
|---|---|
| Endpoints | CLASI-A, CLASI-D, or both; per-sign continuous measures as exploratory endpoints |
| Region set | All 13 regions, photo-exposed sites only, or a sponsor-defined subset |
| Visit structure | Identical capture at every visit, or comprehensive baseline with focused follow-up |
| Target lesions | Number designated at baseline and followed at every visit |
| Area measurement | Relative extent, or absolute area in mm² with marker capture |
| Scale mapping | Fixed mapping from each sign measurement onto its CLASI item scale |
| Colour calibration | Whether calibration markers are included in the capture set |
| Visit schedule | Assessment timepoints aligned with the study calendar |
| Clinician-entered items | Mucosal lesions, recent hair loss, dyspigmentation duration |
| Anonymization | Configured to preserve the facial regions CLASI scores |
| Alert thresholds | Configurable change thresholds for automated notifications |
The scale mapping is fixed for the duration of the study, so that a value reported at the final visit means what the same value meant at baseline.
A study-specific investigator manual is generated for each trial, with step-by-step region capture instructions, example images, and a knowledge test to confirm site training.
2. Site setup
The platform is a web application accessible from any browser, with capture performed on standard smartphones. Each site receives login credentials for study personnel, the pre-configured protocol, the digital investigator manual, and training resources. No hardware installation is required unless the protocol includes calibration markers, which are printed consumables rather than equipment.
3. Patient enrollment
After enrolling a patient per the clinical protocol, the investigator creates the patient record in the platform under a study-specific pseudonymised identifier. This identifier is what binds every subsequent image to a patient without carrying identity.
4. Image capture
The application guides capture region by region, so the region label is applied at the moment of capture rather than reconstructed later. For each region it provides:
- Region guidance showing the required framing and angle
- Real-time DIQA quality check on focus, lighting and framing
- Immediate recapture prompts where an image falls below the quality threshold
- Completion tracking against the protocol's region set for that visit
5. AI scoring
Processing takes approximately <2 seconds per image. The system produces per-region sign measurements, per-sign confidence figures, segmentation overlays showing exactly which pixels drove each extent figure, and a DIQA score per image.
6. Report review
The investigator reviews the assessment in the platform, verifies that the measurements match clinical observation, and completes the components that are not derived from the image: mucous membrane lesions, recent hair loss, and dyspigmentation duration. The report then presents assembled CLASI-A and CLASI-D with the per-region breakdown behind each.
Where an image is inadequate, the investigator can recapture and resubmit before the assessment is finalised.
7. Longitudinal tracking and alerts
For follow-up visits the platform tracks both scores across all assessments:
- CLASI-A evolution with absolute and percentage change from baseline, the usual basis for an activity response endpoint
- CLASI-D accumulation across the study
- Target lesion trajectory for each designated lesion, in absolute area where markers are used
- Repigmentation as reduction in hypopigmented area from baseline, per lesion and per region
- Per-sign trends showing what is driving a change in either score
- Per-region trends showing which regions are responding
- Severity transitions between classification bands
Automated email notifications can be configured against change thresholds, delivered to the site investigator for clinical confirmation and escalation per protocol.
8. Data export and EDC integration
Exported data fields
| Field | Description |
|---|---|
| Patient ID | Pseudonymised study identifier |
| Visit ID | Visit number or scheduled timepoint |
| Timestamp | UTC timestamp of capture and processing |
| Region | CLASI anatomical region |
| Erythema intensity | Per region, on the mapped CLASI item scale |
| Erythema extent | Per region, relative or absolute area |
| Desquamation intensity | Per region |
| Induration intensity | Per region |
| Depigmentation extent | Per region, relative or absolute area |
| Hair loss percentage | Scalp, proportion affected |
| Lesion ID | Target lesion identifier, where designated |
| Lesion area | Absolute area in mm², with marker capture |
| Repigmentation | Reduction in hypopigmented area from baseline |
| CLASI-A | Assembled activity score |
| CLASI-D | Assembled damage score |
| Change from baseline | Absolute and percentage, per score |
| DIQA scores | Per image |
| Confidence figures | Per sign measurement |
Per-sign values are exported alongside the assembled scores rather than only inside them, so that a sponsor can analyse the continuous measures directly. In an early-phase study these carry more resolution than the CLASI items they feed, and they support exploratory analysis of which sign moves first under treatment.
Integration methods
- API-based: Automated data flow from the Legit.Health platform to the sponsor’s EDC system via RESTful API
- CSV export: Structured CSV files for manual import into EDC systems
- CRF field mapping: Data fields pre-mapped to the sponsor’s Case Report Form structure, configured during protocol design
- QuantifiCare platform integration: For studies using the combined QuantifiCare + Legit.Health solution, scoring data flows through QuantifiCare’s platform to the sponsor’s data management system
All exported data is structured for direct mapping to CRF fields in standard EDC systems (Medidata Rave, Veeva Vault CDMS, Oracle InForm, etc.).
The integration is API-based and system-agnostic; it works with any EDC system (Medidata Rave, Oracle InForm, Veeva Vault EDC, and others). Export formats include RESTful API, CSV and Excel, and structured JSON.