ScarBank
Registry schema
Patient, lesion, visit, image, treatment, and outcome tables designed for longitudinal scar research.
AI-native scar phenotyping starter
Open framework for scar and keloid registry workflows.
OpenScar turns longitudinal scar care into structured research assets: a six-table registry schema, synthetic mock data, validation scripts, and clinician-reviewable ScarClaw workflow skills.
Synthetic Case Workflow
Anterior chest keloid, day 0 to day 90
Day 0
Baseline and treatment
Photo capture, symptom intake, and one synthetic steroid injection exposure.
Day 90
Follow-up and review
Structured symptoms, image quality flags, conservative response label, and outcome row.
6registry tables
10ScarClaw skills
5validators/demo scripts
From clinical mess to reusable assets.
The v0.1 starter focuses on standardization before modeling. The goal is to make scar data legible, comparable, and safe to review.
ScarPhoto
Image workflow
Repeatable capture, quality flags, privacy status, and conservative before/after comparison.
ScarTrajectory
Follow-up structure
Symptoms, treatment exposure, scale scores, and outcome labels kept separate and clinician-reviewable.
ScarClaw v0.1 skill pack.
Ten workflow skills cover imaging, clinical follow-up, research abstraction, consent, registry entry, and patient education.
scar-photo-standardizationimaging
scar-image-quality-reviewimaging
before-after-scar-analysisimaging
keloid-follow-up-intakeclinical
scar-treatment-response-reviewclinical
recurrence-risk-factor-extractionresearch
scar-literature-reviewresearch
scar-registry-entry-builderoperations
scar-consent-language-reviewoperations
scar-patient-education-draftpatient care
Minimum viable scar data model.
The schema is small by design so a real clinic can pilot it before any ambitious model training.
patientDe-identified context, skin type, history, consent status.
lesionScar category, site, trigger, duration, baseline symptoms.
visitDate offsets, visit type, symptoms, scale scores.
imageView label, quality flags, scale marker, privacy status.
treatmentModality, details, combination status, adverse events.
outcomeResponse category, recurrence status, confidence, impact.
Retrospective pilot kit.
The next practical asset is a 20-50 case registry workflow that can be run inside an approved clinical environment before any model work.
Protocol
Pilot Registry SOP
Defines cohort boundaries, source material, de-identification flow, review roles, quality gates, and stop conditions.
Abstraction
Case Form
Maps one retrospective case into patient, lesion, visit, image, treatment, and outcome tables.
Templates
CSV Headers
Provides empty schema-compatible CSV files for pilot entry and validation.
Prospective intake pack.
A clinic-facing capture layer turns the stable schema into baseline forms, patient-reported symptom fields, photo metadata, and a REDCap-ready data dictionary.
Clinic workflow
Intake Pack
Defines a prospective capture flow for new visits while keeping OpenScar inside research and documentation support.
Forms
Clinician and Patient Forms
Captures lesion phenotype, symptoms, treatment exposure, outcome windows, and patient-reported concerns.
Data dictionary
Registry Dictionary
Maps every schema field to a prospective source for REDCap, Excel, or internal registry setup.
Safety boundary: OpenScar Starter supports documentation, research preparation,
education, and clinician-reviewable workflows. It does not diagnose, prescribe, determine recurrence,
predict treatment response, or replace licensed clinical judgment.