Longevity, genes,
and disease-modifying AI.

Insilico opens LongevityBench + Longevity Claw (Cell cover)

Open AI longevity stack — domain Longevity-LLMs beat frontier models on aging multi-omics reasoning; Longevity Claw agentically nominated 328 candidate aging targets across 14 hallmarks (enrichment vs known targets; KDM1A independently linked in worm lifespan work).

Caveat: Research/tooling paper + in-silico target nomination — not an approved longevity therapy or clinical efficacy claim; wet-lab and clinical validation still required. (17 Sep 2026.)

PANXEON AI liquid biopsy for early pancreatic cancer (Nature Medicine)

Investigational multi-biomarker + AI blood test detected ~87% of stage I–II PDAC in ~1,800 multinational participants; also flagged high-grade dysplasia in ~64% of evaluated high-risk cysts — potential adjunct to surveillance, not a replacement for imaging.

Caveat: Investigational — not approved for screening/diagnosis; case-control/research setting; prospective utility and outcome benefit not established; false-positive rates higher in high-risk cohorts (~16% vs ~3% low-risk).

IICM+ multimodal AI beats 21-gene Recurrence Score for distant recurrence (TAILORx)

ECOG-ACRIN / Caris model (pathology images + clinical + 42-gene panel) validated in 4,429 TAILORx HR+/HER2− node-negative patients (>11y follow-up) with higher C-index than Oncotype DX for overall, early, and especially late distant recurrence.

Caveat: Prognostic research model only — not cleared to select or change adjuvant therapy; needs external population validation and decision-impact studies. (17 Sep 2026.)

VERVE-102 in vivo PCSK9 base editing — Phase 1 Heart-2 (NEJM / Lilly)

Single-infusion adenine base editor of hepatic PCSK9 in adults with HeFH/premature CAD — reported PCSK9 reductions up to ~88% and LDL-C up to ~62% with signals of durability; supports one-time gene-editing path for severe hypercholesterolemia (not aging).

Caveat: Early-phase investigational gene editing — not approved; sample size limited; infusion reactions / transient ALT elevations / rare serious AE reported in coverage; long-term safety and hard CV outcomes pending.

Research AI ≠ clinical cure claims

Longevity tooling benchmarks, investigational liquid biopsies, prognostic models, and early-phase gene editing are different evidence classes. This site keeps them labeled so a headline never collapses into “AI cured aging.”