πŸ”’ Sign in required

This clinical tool requires sign-in. Your cases are private to your account and are never shared with other users.

GEP-NET Decision & Registry Platform v1.0 Β· prototype

Evidence-graded management support for gastroenteropancreatic neuroendocrine tumours Β· not a substitute for a multidisciplinary tumour board
Assessment
Registry
Follow-up
Evidence base
About & method

1 Β· Tumour & patient inputs

Fields marked * are required to generate grade-dependent recommendations. Optional fields, if left blank, are reported with the clinical consequence of the missing data.

2 Β· Imaging, genetics & fitness

These refine staging and therapy eligibility. Missing entries are flagged with their consequence rather than silently ignored.

Case registry

Every saved assessment is stored locally in this browser and is fully de-identifiable. Use Export to pull a study-ready dataset (CSV or JSON). Click a case to record the management actually delivered and track concordance with the platform's suggestion.
When signed in, your cases sync automatically to your own private registry. Sync now flushes anything queued offline; Load my cases pulls your cases back onto this device. No other user can see your data.

Follow-up capture

Longitudinal follow-up is stored per case for later survival / recurrence analysis. Select a case, then add timepoints.

Embedded evidence base

The decision engine draws on these guideline statements. Level of evidence (LoE) is shown in each guideline's native system: ENETS/NANETS use GRADE (A–D / strong–weak); NCCN uses Category 1 / 2A / 2B; WHO 2022 is a pathology taxonomy (not evidence-graded); trial rows note the study design. Filter by site.

How the engine decides

The platform maps structured inputs to guideline statements and returns every applicable management option, each tagged with its source and level of evidence, sorted by a fixed evidence-priority tier:

Tier 1 Curative surgery

Resection with curative intent when technically/oncologically feasible β€” always ranked first for a surgical candidate.

Tier 2 RCT-backed systemic

Level-I therapies (SSA, PRRT, everolimus, sunitinib) for advanced/unresectable disease.

Tier 3 Locoregional / consensus

Embolisation, surveillance protocols, consensus-based surgical choices.

Tier 4 Ablation (Cat 2B)

Thermal ablation of unresectable oligometastatic disease β€” never above resection for a surgical candidate.

Tier 5 Experimental

EUS-guided ablation and investigational local therapy β€” reserved for patients unfit for/declining surgery, flagged as ungraded.

The core safeguard: if the engine finds a curative surgical option AND a lower-tier ablative/experimental option is theoretically applicable, it demotes the latter and shows an explicit warning β€” directly addressing the problem of experimental RFA being offered ahead of curative dissection.

Storage recommendation

This prototype persists data in the browser (localStorage) with CSV/JSON export β€” zero-setup, single-user, works offline, ideal for evaluation and a single clinician's caseload. For a real multi-centre registry, migrate to a lightweight backend: a hosted Postgres (e.g. Supabase) or REDCap instance behind authentication, with the same field schema used here so exported JSON imports cleanly. That gives multi-user access, audit trails, and IRB-friendly de-identification without changing the data model.

Verification status

Pancreatic and cross-cutting rows were verified against ENETS 2023, NCCN v2.2025, NANETS and WHO 2022. Gastric/duodenal/small-bowel and appendiceal/colorectal rows are encoded from established ENETS/NCCN guidance and should be re-checked against the licensed primary PDFs before clinical deployment. Exact alphanumeric grade labels for a few rows are best-effort and marked accordingly.

GEP-NET Decision & Registry Platform Β· research prototype Β· evidence current to mid-2025 Β· verify against live guidelines before clinical use.
Data stays in your browser. Nothing is transmitted. Export regularly to avoid loss.