Skip to content

Session 3 — Immune-based early cancer detection

Series: 2026 SITC Computational IO Series, session 3 of 10 Date: July 9, 2026, 12:00–1:00 p.m. ET Moderator: Benjamin Vincent, MD — University of North Carolina School of Medicine Faculty: Bo Li, PhD — University of Pennsylvania Recording: SITC OnDemand (registrant / member access) Status: Delivered · content pending recording access

What the session covers

Early cancer detection using immune-repertoire and immune-state readouts rather than tumor-derived signal — i.e. reading the host response instead of, or alongside, ctDNA. Both Li and Vincent work on computational immune-repertoire analysis, so TCR/BCR repertoire inference from bulk and single-cell sequencing is the likely technical core. Inferred from the speakers' fields; confirm against the recording.

Why it matters to this corpus

  • A methodological contrast worth capturing. The corpus's early-detection material leans on tumor-shed signal (ctDNA/MRD, scaffolded in sitc-2026/tools/ and the ASCO 2026 readouts). Host-immune-based detection is a different signal source with different failure modes — a useful axis to have documented.
  • Repertoire inference connects forward to session 9 on AI prediction of TCR–antigen binding, which is the same data type pushed to a harder prediction task.
  • Continues the immunoprevention thread from session 2.
  • Vincent also moderates session 7, tying the early-detection and precision-IO halves of the series together.

Open items

  • [ ] Obtain OnDemand recording; capture the talk title, cohorts, and assay/algorithm details.
  • [ ] Confirm whether the talk covers repertoire inference specifically, or a broader immune-signature panel.
  • [ ] Link the specific tools and papers presented.

Sources