A new method enables robust whole‑genome sequencing (WGS) from formalin‑fixed, paraffin‑embedded (FFPE) specimens, overcoming longstanding technical barriers that limited genomic analysis of archived tumor samples. The technique yields high‑quality genomic data suitable for translational oncology and retrospective studies. In parallel, machine‑learning applied to routine blood counts produced a noninvasive classifier for primary vitreoretinal lymphoma, showing how computational methods can repurpose standard lab data into diagnostic tools. Together these advances lower barriers to molecular diagnosis, expand access to genomic insights from existing tissue banks, and open pathways for earlier, less invasive cancer detection. Pathology labs and clinical genomics services should evaluate validation requirements, data‑quality metrics, and bioinformatics pipelines needed to integrate FFPE‑WGS into diagnostic workflows and trials.