Skip to content

Session 4 — AI, foundation models and medical imaging / radiology

Series: 2026 SITC Computational IO Series, session 4 of 10 Date: July 21, 2026, 12:00–1:00 p.m. ET Moderator: Sandhya Prabhakaran, PhD — Moffitt Cancer Center Faculty: Kun-Hsing Yu, PhD — Harvard University Recording: SITC OnDemand (registrant / member access) Status: Delivered · content pending recording access

What the session covers

Foundation models applied to medical imaging and radiology in an immuno-oncology frame — the series' most direct engagement with the corpus's central AI thread. Yu's group works on machine learning over pathology and multimodal clinical data, so the talk plausibly spans pathology and radiology imaging models rather than radiology alone. Inferred from the speaker's field; confirm against the recording.

Why it matters to this corpus

This is the single highest-priority session in the series for us. It is the point where the corpus's foundation-model spine meets IO biomarker development:

Open items

  • [ ] Prioritize this recording. Obtain OnDemand access and write this page to the SCGD26 standard — thesis, models, findings, Q&A.
  • [ ] Capture which specific foundation models were discussed and on what benchmarks.
  • [ ] Note whether the talk addressed clinical deployment and regulatory posture, which the corpus's imaging vaults track.

Sources