ASCO starts Friday. Here's what 4,773 abstracts reveal about where oncology is headed.

While the five plenaries will dominate the conversation, we analyzed every abstract accepted to ASCO 2026 to surface the cross-cutting themes that tell the real story. The value of a large abstract corpus isn't tracking the trials everyone already knows about. It's seeing where collective research energy is concentrated.

ADCs continue to expand beyond HER2 and TROP2, with 280 abstracts spanning 7 target families. The pipeline is diversifying into B7-H3, c-Met, HER3, and FR-alpha, with earlier-line settings becoming the new competitive frontier.

Bispecifics are crossing from heme into solid tumors. Of 181 bispecific abstracts, 128 are now in solid tumor indications. CD3 T-cell engagers remain the largest class at 79 abstracts, but IO bispecifics are growing fast at 38, with ivonescimab's HARMONi trials anchoring the category.

KRAS druggability is real. 84 agents are in clinical development across 6 mechanism classes, well beyond covalent G12C. Daraxonrasib produced the first Phase 3 survival benefit in PDAC with OS of 13.2 vs 6.7 months. Of the 84 agents, 69 are small molecules, 5 are vaccines, 4 are degraders, and 3 are cell therapies.

GLP-1 receptor agonists are generating a genuine cancer signal. 91 abstracts and a 56-study evidence base show associations between GLP-1 use and reduced cancer incidence, with colorectal showing the strongest protective signal. But only one registered randomized cancer treatment trial exists globally. The signal is real but the evidence remains early.

Liquid biopsy has moved from a technology story to a clinical utility story, with 330 abstracts spanning six use cases from MRD monitoring to early detection. De-escalation is one of the largest themes at the meeting with 632 abstracts, anchored by chemotherapy omission trials like OPTIMA. Cell therapy is expanding beyond CAR-T into TIL, TCR-T, and CAR-NK across 187 abstracts and 122 tracked programs.

The full analysis also covers survivorship and toxicity (543 abstracts), the MCED landscape (28 blood-based detection tests), IO pricing and the pembrolizumab biosimilar cliff, and a forward-looking timeline of catalysts through 2028.