SenoNav AI

A breast cancer diagnosis is only the beginning of a difficult journey. Patients and families are often left with fragmented pathology, imaging and prescription records, unclear biomarker results, missing staging information, and no reliable way to understand which specialist or hospital capabilities they need next. SenoNav AI transforms those scattered records into an evidence-linked Care Passport. It separates confirmed findings from missing, unclear or conflicting information, preserves document and page-level evidence, generates a referral-readiness indicator, maps the capabilities required from the next multidisciplinary care team, and matches hospitals transparently by verified services rather than claiming that any doctor or centre is “the best.” The system is designed for cross-border care journeys, beginning with a Bangladeshi patient preparing for a second opinion in India. It supports patients, families and navigators with plain-language summaries, questions to ask clinicians, document checklists and a downloadable handoff package. SenoNav does not diagnose cancer, prescribe treatment or replace clinicians. Its human-in-the-loop safety layer keeps every recommendation educational, explainable and reviewable. The core inference pipeline was successfully validated on an AMD Instinct MI300X using ROCm, vLLM and Qwen2.5-7B-Instruct. The public demo uses an OpenAI-compatible live endpoint while preserving the same modular architecture for AMD-hosted deployment.