OncoBoard
A patient-facing virtual tumor board where multiple independent AI specialists each review a cancer case, then use consensus for diagnosis questions and majority-vote for treatment-logic questions, surfacing a confidence-rated summary the patient brings to their oncologist.
Recently diagnosed cancer patients and their family members preparing for a second-opinion consultation
- Upload path for PDF reports and pathology notes; an 'AI radiologist', 'AI oncologist', 'AI pathologist', and 'AI GP' each read independently and privately before any exchange
- Adaptive decision mode: consensus pooling for 'what is this?' questions (knowledge-heavy) and majority voting for 'what next?' questions (logic-heavy) — exactly the paper's finding
- Disagreement panel where specialists diverge: the patient sees both rationales in plain language before a final synthesis
- One-page second-opinion brief forwardable to the patient's oncologist, with citations and per-claim confidence
Multi-agent LLM research has matured to a point where consensus vs. voting show clear task-dependent wins — patients shopping for second opinions still wade through PDF overload and a single doctor visit, and an honest, structured 'AI tumor board' hasn't been packaged yet.
Cancer patients regularly seek second opinions, and multiple AI-powered second-opinion services (2ndopinions.ai, continuia.ai, ais2ndopinion.com, secondopinionfromai.com) have launched, confirming a real consumer pull for AI-assisted medical review.2ND Opinion - AI-Powered Medical Opinions ↗Continuia - AI-Powered Second Medical Opinions ↗
Patient-facing AI second opinion is already crowded (at least 4–5 live products), and clinician-facing multi-agent AI tumor boards (ClinicalAI-Board, ml6 AI Tumor Board, the medRxiv VMP paper) already exist; the patient-facing multi-agent-with-consensus-voting combination is a narrow slice, not wide-open.ClinicalAI-Board: AI-powered Virtual Tumor Board ↗The AI-based Tumor Board: A Multi-agent approach ↗
Virtual cancer second opinions already command ~$950 CAD per consult, and patients pay readily for cancer clarity; an AI-augmented product could undercut that while preserving willingness to pay.How Much Is a Private Medical Consultation? ↗OpinionMD Second Opinion Service ↗
Cancer incidence and second-opinion demand are durable, and multi-agent LLM consensus/voting is a lasting architectural pattern, though FDA SaMD risk and incumbent replication could compress the moat.FDA Issues Comprehensive Draft Guidance for AI-Enabled Medical Devices ↗
A patient-facing tool that claims to help diagnose or guide cancer treatment likely triggers FDA SaMD classification; medical liability for misdiagnosis is severe, and the multi-agent consensus logic itself still has limited clinical validation beyond early medRxiv work.FDA Oversight: Understanding the Regulation of Health AI Tools ↗AI-driven virtual tumor board enhances precision (medRxiv) ↗