Upload symptoms and follow patterns
The prototype direction centers on structured symptom data, trend timelines, and clearer summaries patients could bring into physician visits.
A Gastro Compass project
GastroLens app available
Meet GastroLens, the interactive research-mode prototype from GastroCompass.
Gastro Compass is the umbrella project building GastroLens. GastroLens is the upcoming research-mode prototype built to organize persistent symptom patterns, risk context, and follow-up-ready summaries for early-onset colorectal cancer concern. It is non-diagnostic, not clinically validated, and now connected to the main site as the interactive demo layer.
Start with a sample scenario
Minimal demo
Persistent bleeding
High-cluster demo
These presets mirror the real GastroLens demo and judging flow.
What GastroLens is built around
Structured intake
Explainable scoring
Research-mode results
A step-based intake leads to an explainable research summary.
Research mode, not diagnosis
Informational only
Non-diagnostic
Clinician follow-up encouraged
Safety language stays non-diagnostic and follow-up oriented.
Core symptoms
Rectal bleeding
Persistent
Bowel change
Significant
Duration
6+ weeks
Built as a short guided intake rather than a long medical questionnaire.
Move from symptoms to summary
1
Core symptoms
Current step
2
Additional signals
Next up
3
Risk context
Family history and screening
4
Review and generate
Prepare summary
The structure is visual, quick to complete, and easy to explain in a demo.
Guided from start to result
Sample scenarios
Guided prompts
Generate summary
The preview mirrors the real GastroLens flow without pretending the product is public yet.
Moderate flag
Total score 8
Pattern lens: persistent bleeding
Driver summary: duration + bowel change
Follow-up-oriented language
Based on the real GastroLens results panel.
Core symptoms
4 points
Additional pattern
2 points
Risk context
2 points
Scores are broken into explainable sections.
Built for escalation support
Pattern explanation
Safety note
Care conversation
The result is framed as a reason to follow up, not a diagnosis.
What Judge Mode reveals
Interaction bonuses
Shows weighted rule interactions behind the score.
Validation notes
Flags contradictions and edge cases in the logic.
Presentation actions
Copy summary
Print summary
Start a new assessment
Designed for demos, judging, and review sessions.
Deterministic, explainable, and demo-ready
Step-based intake
Transparent scoring
Judge mode extras
This preview now reflects the actual GastroLens app structure.
Prototype architecture
How the GastroLens concept is being structured
Patient View
Symptom and context intake
Rectal bleeding, bowel changes, abdominal pain, fatigue, duration, family history, age, ethnicity, and prior colonoscopy status.
GastroLens prototype
Symptom-first risk flagging
Structured analysis of persistence, recurrence, and documented risk context, with imaging review kept as a future validated layer.
Professional View
Action-ready review
Clinician-facing summaries designed to support earlier follow-up conversations, not replace clinical judgment.
Preview views
One symptom-first system, with imaging kept as roadmap work
Future imaging review
The imaging component remains future work until it can be trained and validated on a clinically meaningful dataset with clear performance thresholds.
GastroLens workflow preview
What the upcoming workflow is designed to do
01
Capture
Patients would upload symptoms through structured prompts along with risk context such as age, ethnicity, family history, and prior colonoscopy status.
- Patient symptom timeline
- Age and demographic intake
- Family history and prior screening
- Context and notes
02
Organize
The prototype converts daily logs into a structured timeline so patterns are visible instead of buried across isolated notes.
6-week persistence window
03
Analyze
The planned first model would look for symptom clusters, persistence, and risk stratification patterns that support earlier follow-up in the right patients.
Symptom classifier
Risk context
Duration weighting
Escalation support
04
Escalate
The result is not a diagnosis. It is a clearer signal that says the pattern warrants a conversation with a physician.
Follow-up support
Future summaries are intended to highlight persistent patterns that should not be dismissed without review.
Shared output
Built to make the next decision clearer
GastroLens is being organized to turn scattered symptom history and risk context into one usable output: a planned AI-supported review that is easier to act on and easier to bring into care.
Clearer symptom history
Patients get organized timelines and summaries they can understand and share.
Escalation-ready summary
Professionals get organized symptom context, risk factors, and one clearer review pathway for follow-up conversations.
Research foundation
Why GastroLens is being built this way
GastroLens is coming out of a research-first process, not a rushed app launch. The work behind it includes the completed research paper, the prototype design process, NHANES-grounded feasibility work, and three empirical analyses used to test whether the concept is credible enough to build. That is why the prototype is being kept clearly non-diagnostic and clearly in development.
A preview-stage prototype grounded in evidence, shaped by design work, and held to a safety-first standard before public release.
Coming soon