Scan analysis and review
From 3D scan scoring to practice-level review
PitchAI developed a dental scan-analysis pipeline, then extended the work into dashboards with practice-scoped access and organization comparisons.
- Context
- Clinical workflows
- Stage
- Deployed platform
- Published
Begin with the scan calculation
The initial scope covered classifying and segmenting 3D dental scans, identifying measurement points and calculating a weighted score from five components.
The implementation processes component results separately before combining them. It records component scores alongside their weighted values. The aggregation step checks for unknown or duplicate components, missing results and invalid score values.
A full total is available only when every required component has a score. If a component is unavailable, the total remains unavailable too.
Keep quality indicators with the case
The processing workflow also records potential scan-alignment and inference-quality issues. Those indicators are saved on the case, separately from the scoring result.
A flag is not an automatic rejection. Scoring can still run after an indicator is recorded. The implementation distinguishes a successful calculation and its report from a failed calculation with an error record.
The calculation and its quality indicators therefore remain separate records for review.
Extend the work into practice dashboards
A later phase added integration and reporting work around the scoring pipeline. The web dashboards include practice-level tables, trends and comparisons with organization aggregates.
Practice users receive their own practice view. The comparison is an anonymous organization benchmark rather than a list of named peer practices. Authorized internal users retain the broader analysis view.
A practice can review its selected cases against the organization comparison without receiving the underlying named peer rows.
Enforce the scope before returning data
The dashboard resolves a practice user's scope from the signed-in session. A request cannot switch that scope by supplying another practice identifier.
After fetching practice rows, a second check rejects a response if it contains a practice outside that scope. The practice-facing table labels its own row and presents the organization comparison as an aggregate.
The platform was deployed, and the practice-dashboard access work has production verification from July 2026. That review covered rendered pages, response payloads and attempts to open another practice's views.
The initial scoring work grew into a review workflow with component records, quality indicators and controlled access to comparisons.