A complete review—not just an AI estimate
See the original model output, the evidence available to the clinician, the completed feedback, the Save action, and the persisted reviewed result.
Product demo · clinician review workflow
See how a hand radiograph becomes a clinician-reviewed bone-age result.
This guided demo follows the complete workflow: upload a synthetic hand-and-wrist radiograph, run the analysis, understand the estimate and supporting evidence, record the clinician’s decision, and confirm that the review was saved.
Watch the Bone Age Calculator workflow
In 39 seconds, follow six numbered steps from image selection to a saved clinician review. Blue click markers show each action, while plain-language captions explain what the outputs mean and why the clinician remains in control.
Prefer a continuous screen recording?
Watch the same six-step workflow as one uninterrupted browser session, including the real typing, clicks, analysis, review, and saved-result confirmation.
Need a quick preview?
Scan the key moments in a short silent animation, then use the guided video above for the complete workflow and output explanations.
How the workflow works
The calculator begins with secure account access. This demo starts inside the clinician workspace and uses a synthetic image and synthetic values, so no patient data appears anywhere in the walkthrough.
Prepare the case
Choose one clear hand-and-wrist radiograph and confirm the sex used by the calculator. The example image is synthetic and contains no patient data.
Analyze the image
Select Analyze radiograph. At a high level, the calculator reads visible growth patterns across the hand and wrist and compares them with labeled examples learned during model development.
Understand the outputs
Read the estimated age, range, and confidence together, then compare the ranked reference images. These are decision-support outputs, not a diagnosis; their meanings are explained below.
Record the clinician decision
Accept the estimate, choose a reference, enter a corrected age with a reason, or mark the image unusable. The demonstration records a corrected age of 124 months.
Save the reviewed result
Select Save reviewed result. The clinician decision and reason are stored separately so they do not overwrite the original model result.
Verify the record
Reload the page and confirm the saved decision is read back while the original estimate remains unchanged. This is the visible proof that the review was stored durably.
What the outputs mean
The calculator summarizes patterns visible in the radiograph. It does not determine a diagnosis, and the clinician remains responsible for the final review.
The estimated age is the model’s best summary of skeletal maturity, expressed in months and years. The range shows a nearby interval around that estimate, reflecting that image-based estimates are not exact.
Confidence summarizes how strongly the model supports its estimate. The ranked references are visually similar examples with known ages, provided so the clinician can compare the visible growth patterns.
The clinician can accept, correct, or reject the model result and add a reason. This human decision is saved separately from the model estimate so both remain understandable and traceable.
What success looks like
The current radiograph, three ranked references, estimated age, range, confidence, and disagreement reason appear in the product UI.
No review option is selected automatically. The clinician explicitly chooses and saves the decision.
After navigation reload, the saved decision is read from the API and the immutable model evidence remains separate.
Safety and scope
Research demonstration—not a medical device or clinical diagnosis.
The video uses a programmatically generated radiograph, a deterministic mock inference provider and a temporary local durable store. It proves the product workflow, not diagnostic accuracy. The public app may run a different model revision; independently validate performance, acquisition quality, governance and clinical suitability before any real-world use.