GAL-001: Adaptive Phase II in Refractory Melanoma

active NXC-101 (anti-PD-1 checkpoint inhibitor)

Phase II · Refractory melanoma · Sponsor: Galen Therapeutics

Agentic orchestration Horizon 3

Runs all twelve lifecycle steps below in order; each step reports what it did and escalates to a human only on exception.

Study lifecycle checklist

Twelve steps in trial order. Expand a step to see exactly what the agent did and what it produced; steps needing a decision link to the review queue.

via pipeline

Protocols (1)

v0.1-auto pending_review

Feasibility (2)

US — score 0.45 (12 sites, 90 patients, 10 mo)

At risk: 0.75 patients/site/month (1.0 after projection markdown) strains typical capacity. Human review required before committing timelines. Termination-risk factors: phase II (elevated early-termination odds).

EU — score 0.0 (3 sites, 240 patients, 6 mo)

At risk: 13.33 patients/site/month (17.78 after projection markdown) strains typical capacity. Human review required before committing timelines. Termination-risk factors: phase II (elevated early-termination odds).

Supply forecasts (2)

6-month horizon: 7 kits (7 predicted patients × 0.8 retention, 15% buffer)

Regulatory submissions (1)

FDA · IND amendment pending_review

Study documents 5 files

Everything generated or filed for this study. Click a file to preview it as a PDF.

Agent reasoning trail 13 runs

Explainable, traceable, inspection-ready — every automated action records why it did what it did (FDA/EMA January 2026 AI guiding principles).

Agent	Horizon	Outcome	Reasoning
feasibility_agent	horizon_1	skipped	Feasibility assessment already on file; nothing to hand off.
protocol_agent	horizon_1	skipped	A protocol version already exists; drafting again would fork the design.
site_activation_agent	horizon_2	completed	All 3 sites pass the essential-document checklist.
recruitment_agent	horizon_2	escalated	Queried partner matching engine: 14 candidates. Patient-facing step, so import is human-gated.
monitoring_agent	horizon_2	completed	Risk-based monitoring pass complete — Mount Halifax Oncology Center: 0.1, St. Brigid University Hospital: 0.2, Pacific Trials Institute: 0.0. No site above threshold.
cdm_agent	horizon_1	completed	CDM sweep: generated 0 queries from consistency rules; auto-resolved 0; routed 0 to a data manager. No query past the 14-day JSCDM SLA.
safety_agent	horizon_2	skipped	No unprocessed adverse events; pharmacovigilance queue is clear.
retention_agent	horizon_2	escalated	Scored 4 active patient(s); 3 at elevated dropout risk. Outreach proposed (patient contact is human-gated).
termination_watch_agent	horizon_2	escalated	Termination-risk surveillance raised 3 flag(s): Accrual pace 11% of the pro-rata plan (4 accrued vs ~36 expected; heuristic threshold 60% — poor accrual causes 34.5% of early terminations) · Phase II (elevated early-termination odds, OR 1.6–2.4) · Serious-AE density 33% (1 serious AEs / 3 enrolled). Termination is a human decision — only efficacy, safety, or feasibility justify it (Lièvre 2001).
supply_agent	partner	completed	Supply model handed to partner:irt-rtsm: 7 kits over 6 months (retention 0.8).
biostats_agent	horizon_1	escalated	Generated 2 TLF outputs from live data; batch parked for validation.
medical_writing_agent	horizon_1	skipped	A CSR draft is already awaiting medical writing review.
tmf_agent	horizon_1	completed	TMF self-audit: 1.0 completeness — inspection-ready.