A MIT-led approach validated buoyant mass of T cells as a predictor of patient response to immune checkpoint blockade in melanoma. In a clinical study of 28 patients, buoyant mass outperformed tumor mutational burden as a responsiveness signal, positioning a physical cell-state metric as a potential alternative or complement to molecular biomarkers. In Alzheimer’s imaging triage, an externally validated two-stage blood-biomarker algorithm reduced the need for amyloid PET by resolving uncertainty from p-Tau217. The workflow keeps PET for residual uncertainty while potentially lowering costs and burden. Both developments show a common direction: shifting from single-test biomarkers toward measurement frameworks that manage uncertainty and better predict downstream clinical response.